Understanding the shape properties of trihedral polyhedra
- 格式:pdf
- 大小:217.00 KB
- 文档页数:11
双曲线题型解题技巧The hyperbola is a type of conic section that can be described by the equation (x-h)^2/a^2 - (y-k)^2/b^2 = 1. It has two branches that open in opposite directions and intersect at a point called the center of the hyperbola. The hyperbola is a versatile curve that appears in many real-world applications, such as satellite orbits, antennas, and cooling towers.双曲线是一种可以由方程(x-h)^2/a^2 - (y-k)^2/b^2 = 1描述的圆锥曲线。
它有两个分支,分别朝着相反的方向打开,并在一个称为双曲线中心的点相交。
双曲线是一种多才多艺的曲线,在许多现实世界的应用中都可以看到,比如卫星轨道、天线和冷却塔。
One key technique for solving problems involving hyperbolas is understanding their properties, such as the asymptotes that guidethe shape of the curve. The asymptotes are imaginary lines that the hyperbola approaches but never touches, and they can help us visualize and analyze the behavior of the curve. By studying the asymptotes, we can gain insights into the general behavior of the hyperbola and make solving problems involving them much easier.解决涉及双曲线问题的一个关键技巧是了解它们的性质,比如指导曲线形状的渐近线。
园林英语知识点总结In this article, we will cover a wide range of gardening knowledge points, including plant selection, soil and fertilization, watering, pruning, and pest control. Whether you are a beginner or an experienced gardener, this comprehensive guide will provide you with the essential information and tips to help you create and maintain a beautiful and healthy garden.1. Plant SelectionSelecting the right plants for your garden is crucial to its success. Different plants have different requirements for sunlight, soil, and water, so it is essential to choose plants that are suited to your garden's specific conditions. Consider factors such as the climate, soil type, and available space when selecting plants for your garden.When choosing plants, consider their growth habit, size, and maintenance requirements. Annuals, biennials, and perennials are the three main types of plants, each with its unique characteristics and life cycles. Annuals complete their life cycle in one season, while biennials take two years to complete their life cycle, and perennials live for multiple years. Understanding these differences will help you plan and design your garden effectively.2. Soil and FertilizationSoil is the foundation of a healthy garden, and understanding the properties of your soil can help you improve its fertility and structure. Conduct a soil test to determine the pH level, nutrient content, and texture of your soil. This information will guide you in choosing the right fertilizers and soil amendments to improve the quality of your soil.Fertilizing your garden is essential to ensure that your plants receive the nutrients they need to thrive. There are various types of fertilizers, including organic and synthetic options, each with its advantages and disadvantages. It is essential to follow the recommendations on the fertilizer label and apply it at the right time and in the right amount to avoid over-fertilization, which can harm your plants and the environment.3. WateringProper watering is critical to the success of your garden. Different plants have different water requirements, so it is essential to understand the needs of your plants and water them accordingly. Factors such as the type of plant, soil type, weather conditions, and time of year will all affect how much and how often you should water your garden.Watering your garden in the morning is generally the best practice, as it allows the plants to absorb the water before the sun's heat causes it to evaporate. Mulching your garden can also help retain moisture in the soil, reduce water evaporation, and prevent the growth of weeds. Regularly monitoring the moisture level in the soil will help you determine when and how much to water your garden.4. PruningPruning is an essential gardening practice that helps maintain the health, appearance, and productivity of your plants. Pruning removes dead or diseased branches, shapes the plant, and promotes new growth. Different plants have different pruning needs, so it is essential to understand the pruning requirements of your plants to ensure that you prune them effectively.When pruning, use sharp, clean tools to make clean cuts and avoid damaging the plant. Different pruning techniques, including pinching, thinning, and heading, can be used to achieve different goals, such as controlling the size and shape of the plant, promoting flowering, or enhancing the plant's overall health. Prune your plants at the right time and in the right way to maintain their vitality and beauty.5. Pest ControlPest control is essential to protect your garden from damage caused by insects, diseases, and other pests. There are various methods of pest control, including cultural, biological, and chemical control options. Cultural control involves practices such as crop rotation, proper plant spacing, and sanitation to prevent pest infestations. Biological control involves the use of natural enemies, such as predators and parasites, to control pest populations. Chemical control, such as the use of insecticides and fungicides, is often used as a last resort when other methods have failed to control pest problems. When using chemical control, it is essential to carefully follow the label instructions, use the product safely and responsibly, and consider the potential impact on beneficial insects and the environment.6. Garden MaintenanceRegular maintenance is essential to keep your garden healthy and beautiful. This includes tasks such as weeding, deadheading, and dividing plants, as well as monitoring for signs of pests and diseases. Removing weeds helps reduce competition for water, nutrients, and sunlight, and prevents them from spreading and taking over your garden. Deadheading, or removing spent flowers, helps promote continuous blooming and prevents the plant from producing seeds. Dividing plants, such as perennials, can help rejuvenate them, control their size, and propagate them to create new plants. Additionally, monitoring for signs of pests, diseases, and other issues will help you address problems early and prevent them from spreading and causing significant damage to your garden.7. Garden DesignDesigning your garden involves planning and arranging plants, paths, and structures to create a beautiful and functional outdoor space. Factors such as the layout, style, and theme of your garden will influence its overall design. Consider the architectural style of your home, the size and shape of your garden, and your personal preferences when designing your garden.When designing your garden, consider elements such as color, texture, and form to create visual interest and harmony. Pay attention to the principles of design, such as balance, proportion, and focal points, to create a well-organized and visually appealing garden. Paths, fences, walls, and other hardscape elements can also enhance the functionality and aesthetics of your garden.8. SustainabilityCreating a sustainable garden involves using environmentally friendly practices to conserve natural resources, reduce waste, and minimize the impact on the environment. This includes practices such as reducing water consumption, using organic fertilizers and pesticides, and attracting beneficial wildlife to your garden.Water-efficient gardening practices, such as xeriscaping, mulching, and rainwater harvesting, can help you reduce water usage and create a more environmentally friendly garden. Using organic fertilizers and pesticides, such as compost and biological control agents, can help protect the environment and the health of beneficial insects. Additionally, creating habitat for beneficial wildlife, such as birds, butterflies, and bees, can help promote ecological balance and biodiversity in your garden.In conclusion, gardening is a rewarding and enjoyable activity that allows you to connect with nature and create a beautiful outdoor space. By understanding the essential knowledge points of gardening, such as plant selection, soil and fertilization, watering, pruning, pest control, garden maintenance, design, and sustainability, you can create and maintain a thriving garden that brings you joy and satisfaction. Whether you are a beginner or an experienced gardener, this comprehensive guide will provide you with the information and tips you need to succeed in your gardening endeavors.。
英语语音有关知识的作文Title: Exploring the Fundamentals of English Phonetics。
English phonetics, the study of the sounds of the English language, is a fascinating area of linguistics that plays a crucial role in understanding and effectively communicating in English. In this essay, we will delve into the key aspects of English phonetics, including its importance, basic concepts, and practical applications.First and foremost, the significance of English phonetics cannot be overstated. It serves as the foundation for accurate pronunciation, clear communication, and effective language learning. Mastery of English phonetics enables learners to produce and recognize the diversesounds of English, thereby enhancing their listening, speaking, and even reading skills.At the heart of English phonetics lie its consonantsand vowels, the building blocks of spoken language.Consonants are sounds produced by obstructing orrestricting the airflow in the vocal tract, while vowelsare produced with an open vocal tract. Understanding the articulatory properties of consonants and vowels isessential for accurately reproducing them in speech.One fundamental concept in English phonetics is the International Phonetic Alphabet (IPA), a standardizedsystem of phonetic notation that represents the sounds of spoken language. The IPA provides a precise and consistent way to transcribe the sounds of English, allowing linguists, language learners, and educators to describe and analyze pronunciation with clarity and accuracy.In English, consonants can be classified based on various articulatory features, such as place ofarticulation, manner of articulation, and voicing. For example, consonants like /p/, /t/, and /k/ are producedwith a complete closure or blockage of airflow (stops), while consonants like /s/, /z/, and /ʃ/ involve a partial obstruction of airflow (fricatives). Understanding these articulatory distinctions helps learners distinguishbetween similar sounds and improve their pronunciation.Vowels in English are characterized by their positionin the mouth and the shape of the tongue. The English vowel system consists of short vowels (e.g., /æ/, /ɛ/, /ɪ/) and long vowels (e.g., /iː/, /uː/, /ɔː/), as well as diphthongs (e.g., /eɪ/, /aɪ/, /ɔɪ/), which are combinations of two vowel sounds within the same syllable. Mastering the pronunciation of vowels is crucial for achieving intelligible speech and avoiding misunderstandings.Furthermore, English phonetics encompasses suprasegmental features, such as stress, intonation, and rhythm, which contribute to the prosody of spoken English. Stress refers to the emphasis placed on certain syllables within words, while intonation involves variations in pitch that convey meaning and attitude. Rhythm refers to the pattern of stressed and unstressed syllables in connected speech, which influences the overall flow and naturalness of spoken language.Practical applications of English phonetics abound in everyday communication, language teaching, and speech technology. Language learners can benefit from phonetic training to improve their pronunciation and oral proficiency. Teachers can use phonetic transcription to explain pronunciation patterns and assist students in developing their listening and speaking skills. Speech recognition systems rely on phonetic algorithms to accurately interpret and process spoken input, enabling applications such as voice search, virtual assistants, and language learning platforms.In conclusion, English phonetics plays a vital role in understanding the sounds of English, enhancing communication, and facilitating language learning. By mastering the phonetic aspects of English, learners can achieve clearer pronunciation, better comprehension, and more effective communication in both spoken and written contexts. With its practical applications and theoretical insights, English phonetics remains an indispensable field of study for anyone interested in the English language and its diverse linguistic features.。
conformal基本流程英文回答:Conformal Flow.Conformal flow is a powerful technique in manifold learning that helps in understanding the intrinsicstructure of data lying on a manifold. It is an iterative process that gradually deforms a given initial shape (e.g., a sphere) to match the underlying data distribution. The deformation is guided by a metric that measures the discrepancy between the shape and the data.Conformal Flow Algorithm.The conformal flow algorithm involves the following steps:Initialization: Start with an initial shape, typically a sphere, representing the manifold.Embedding: Embed the data points into the initialshape using a suitable embedding technique.Metric Computation: Calculate the discrepancy metric between the embedded data and the shape.Deformation: Deform the shape by updating its vertices based on the gradient of the discrepancy metric.Iteration: Repeat steps 2-4 until the discrepancymetric converges or a desired level of accuracy is achieved.Regularization:To prevent overfitting, regularization techniques are employed in conformal flow. Common regularization methods include:Smoothing: Smoothing the deformed shape to reducenoise and improve stability.Landmark Constraints: Fixing certain landmarks on the shape to maintain the overall structure.Penalizing Distortion: Penalizing excessive deformations to prevent the shape from collapsing.Applications:Conformal flow has been successfully applied in various domains, including:Image Segmentation: Identifying and segmenting objects in images.Shape Analysis: Analyzing the shape and properties of 3D objects.Data Visualization: Visualizing high-dimensional data by projecting it onto a lower-dimensional manifold.Machine Learning: Improving the performance of machine learning models by understanding the underlying datastructure.中文回答:共形流。
介绍几何图形的作文英语Geometry is a branch of mathematics that deals with the study of shapes, sizes, and positions of objects in space.It is an important subject that has many practical applications in our daily lives. In this essay, I will introduce several common geometric shapes and their properties.The first shape I would like to introduce is the circle.A circle is a closed curve that is made up of points that are all the same distance from a fixed point called the center. The distance from the center to any point on the circle is called the radius. The circumference of a circleis the distance around the outside edge, and it can be calculated by multiplying the diameter (the distance across the circle through the center) by pi (approximately 3.14). Circles are found in many everyday objects, such as wheels, plates, and coins.Another common geometric shape is the triangle. Atriangle is a three-sided polygon that has three angles. The sum of the angles in a triangle is always 180 degrees. Triangles can be classified based on the length of their sides and the size of their angles. For example, an equilateral triangle has three equal sides and three equal angles of 60 degrees each, while a right triangle has one angle that is 90 degrees. Triangles are used in construction to create stable structures, such as roofs and bridges.Rectangles are another important geometric shape. A rectangle is a four-sided polygon that has two pairs of parallel sides and four right angles. The length of a rectangle is its longer side, while the width is its shorter side. The area of a rectangle can be calculated by multiplying its length by its width, while its perimeter (the distance around the outside edge) is the sum of its four sides. Rectangles are commonly used in architecture and design to create buildings and furniture.Finally, I would like to introduce the sphere. A sphere is a three-dimensional shape that is formed by rotating acircle around its diameter. It is a perfectly round shape that has no edges or corners. The surface area of a sphere can be calculated by multiplying its radius by 4 pi, while its volume can be calculated by multiplying the cube of its radius by 4/3 pi. Spheres are found in many natural objects, such as planets and bubbles.In conclusion, geometry is a fascinating subject that has many practical applications in our daily lives. By understanding the properties of common geometric shapes, we can better appreciate the beauty and complexity of theworld around us.。
英语语音概念知识点总结Phonetics is a branch of linguistics that studies the sounds of human speech. It is concerned with the physical properties of speech sounds, how they are produced and perceived, and how they function in language. English phonetics specifically focuses on the sounds of the English language and how they are used in communication.There are three main aspects of phonetics: articulatory phonetics, acoustic phonetics, and auditory phonetics. Articulatory phonetics is concerned with how speech sounds are produced by the articulatory organs, such as the lips, tongue, and vocal cords. Acoustic phonetics studies the physical properties of speech sounds, such as their frequency and intensity, and how they are transmitted through the air. Auditory phonetics deals with how speech sounds are perceived by the ear and interpreted by the brain.Articulatory PhoneticsArticulatory phonetics is concerned with the physical movements and configurations of the speech organs that produce speech sounds. These organs include the lips, tongue, teeth, palate, and vocal cords. The study of articulatory phonetics involves understanding how these organs come together to produce specific speech sounds.In English, there are several important articulatory features that contribute to speech sounds. These include the position of the tongue, the shape of the lips, and the opening and closing of the vocal cords. For example, the /p/ sound is produced with the lips closed and then suddenly released (bilabial plosive), while the /s/ sound is produced with the tongue near the alveolar ridge and air passing through a narrow channel (alveolar fricative). Understanding the articulatory features of speech sounds is essential for language learners and speech therapists, as it helps them to produce and analyze speech sounds accurately. For example, knowing the correct tongue position for the /r/ sound can help non-native speakers improve their pronunciation, and understanding how the vocal cords are used for voiced sounds can assist in diagnosing speech disorders.Acoustic PhoneticsAcoustic phonetics is concerned with the physical properties of speech sounds, such as their frequency, amplitude, and duration. These properties are measured using instruments such as spectrograms, which can graphically represent the acoustic characteristics of speech sounds.In English, there are specific acoustic properties that distinguish the different speech sounds. Vowels, for example, are characterized by their formants, which are specific frequency ranges that are amplified in the sound spectrum. Consonants, on the other hand, can be characterized by their burst, which is a sudden release of air, or by their noise, which is aperiodic sound caused by turbulence.Understanding the acoustic properties of speech sounds is important for speech recognition technology and language assessment. For example, automatic speech recognition systems use acoustic models to understand and transcribe spoken language, and speech-language pathologists use acoustic analysis to assess and diagnose speech disorders.Auditory PhoneticsAuditory phonetics is concerned with how speech sounds are perceived by the ear and processed by the brain. It involves understanding how the ear detects sound waves, how the brain interprets these signals, and how this information is used in language processing.In English, there are specific auditory features that help to distinguish speech sounds. For example, vowels are perceived based on their pitch and timbre, while consonants are perceived based on their characteristic noise or release. The brain processes these auditory signals to recognize and understand words and sentences in spoken language.Understanding the auditory perception of speech sounds is important for language learners and educators. For example, knowing how the ear detects and processes speech sounds can help language learners improve their listening skills, and understanding how the brain interprets speech sounds can assist educators in designing effective language instruction. Phonetic TranscriptionPhonetic transcription is a system for representing speech sounds using symbols. In English, the International Phonetic Alphabet (IPA) is commonly used for phonetic transcription. The IPA provides a set of symbols that represent the sounds of all human languages, including the sounds of English.Phonetic transcription is used for a variety of purposes, including language teaching, speech pathology, and linguistic research. For example, language teachers use phonetic transcription to help students learn the correct pronunciation of words, speech pathologists use it to diagnose and treat speech disorders, and linguists use it to analyze the sounds of different languages.In phonetic transcription, each speech sound is represented by a specific symbol that corresponds to its articulatory, acoustic, and auditory properties. For example, the /s/ sound is represented by the symbol [s] in phonetic transcription, while the /p/ sound is represented by the symbol [p]. These symbols allow linguists and language learners to accurately represent and analyze the sounds of English.ConclusionEnglish phonetics is a complex and fascinating field that plays a crucial role in the study of language and communication. By understanding how speech sounds are produced, how they are transmitted through the air, and how they are perceived by the ear, we can gain valuable insights into the nature of human speech and the ways in which it functions in language. Phonetic transcription provides a valuable tool for representing and analyzingspeech sounds, and it is used in a variety of fields, including language teaching, speech pathology, and linguistic research. Overall, English phonetics is an essential component of the study of language and an important area of research for linguists, educators, and speech professionals.。
离心率公式双曲线The eccentricity formula for a hyperbola is a fundamental concept in mathematics that plays a crucial role in understanding the shape and properties of hyperbolas. It defines the amount by which a hyperbola deviates from being circular and is a key parameter that can be used to describe the geometry of the hyperbola.双曲线的离心率可以通过以下公式来计算:e = √(a^2 + b^2)/a,其中a 和b分别是双曲线的两个轴的长度。
离心率是一个0到1之间的值,当离心率接近于0时,双曲线形状接近于椭圆;当离心率接近于1时,双曲线形状变得非常扁平,曲线变得非常陡峭。
这个公式是双曲线性质的重要指标,可以帮助我们更好地理解双曲线的形状和特性。
Understanding the eccentricity formula for hyperbolas also has practical applications in various fields such as physics, engineering, and astronomy. For example, in physics, the eccentricity of an orbit can determine the shape and behavior of planetary trajectories. In engineering, the eccentricity of structures like bridges or arches can affect their stability and load-bearing capacity. In astronomy, theeccentricity of comets' orbits can indicate how elongated their paths are around the sun.双曲线的离心率公式还可以在教育领域得到应用,通过教授学生如何计算和理解离心率,可以帮助他们更好地掌握几何学的概念,提高数学素养。
宗师级铸锤任务流程As a blacksmith, becoming a grandmaster in the craft of forging hammers is an esteemed accomplishment that requires dedication, skill, and perseverance. The journey to achieving this prestigious title is a rigorous and challenging one that tests both your technical abilities and creative prowess. From mastering traditional techniques to pushing the boundaries of innovation, the path to becoming a grandmaster in hammer forging is not for the faint of heart.作为一名铁匠,在锻造铸锤匠宗师这一领域取得成就是一个值得尊敬的成就,它需要全身心的投入、技术技巧和坚持不懈的努力。
要获得这一崇高头衔的旅程是一个严格而充满挑战的过程,它考验着你的技术能力和创造力。
从掌握传统技巧到推动创新的边界,成为一名锻造铸锤匠宗师的道路并不适合心志薄弱者。
One of the first steps in the journey towards becoming a grandmaster in hammer forging is mastering the fundamentals of blacksmithing. This includes learning how to properly heat and shape metal, understanding the properties of different metals, and honing your precision and attention to detail. Without a strong foundationin basic blacksmithing skills, it is nearly impossible to progress to the level of grandmaster in hammer forging.在成为铸锤匠宗师之路上的第一步是掌握锻造基础知识。
土木工程材料英语Civil Engineering Materials。
Introduction:Civil engineering materials are the cornerstone of any construction project. They play a crucial role in determining the strength, durability, and overall performance of structures. In this article, we will explore the various types of materials used in civil engineering and their properties.Concrete:Concrete is the most widely used construction material in civil engineering. It is composed of cement, fine aggregates (sand), coarse aggregates (gravel or crushed stones), and water. The chemical reaction between cement and water, known as hydration, results in the formation of a solid mass that binds the aggregates together. Concrete exhibits excellent compressive strength but has relatively low tensile strength. To overcome this limitation, reinforcement in the form of steel bars is added to create reinforced concrete.Steel:Steel is another essential material in civil engineering. It is primarily used for structural purposes due to its high strength and ductility. Steel structures can withstand heavy loads and are resistant to earthquakes and other external forces. The two main types of steel used in construction are carbon steel and stainless steel. Carbon steel is more commonly used due to its cost-effectiveness, while stainless steel offers superior corrosion resistance.Brick:Bricks have been used in construction for thousands of years and are still widely used today. They are made by firing clay or shale in a kiln. Bricks are known for their durability, thermal insulation properties, and aesthetic appeal. They are commonly usedin the construction of walls, pavements, and arches. However, bricks have low tensilestrength and are susceptible to weathering, which can affect their structural integrity over time.Timber:Timber is a renewable resource that has been used in construction for centuries. It is lightweight, easy to work with, and exhibits good thermal insulation properties. Timber structures are commonly found in residential buildings, bridges, and temporary structures. However, timber is susceptible to decay, insect infestation, and fire. To enhance its durability, timber can be treated with preservatives.Asphalt:Asphalt, also known as bitumen, is widely used as a paving material in road construction. It is a mixture of aggregates, such as sand and gravel, and bitumen, a sticky, black, and highly viscous substance. Asphalt pavements offer excellent durability, skid resistance, and noise reduction. They can withstand heavy traffic loads and are relatively low in cost compared to other paving materials.Aggregate:Aggregates are essential components in concrete and asphalt mixtures. They provide bulk to the mixture and contribute to its strength and stability. The main types of aggregates used in civil engineering include sand, gravel, crushed stone, and recycled concrete. The properties of aggregates, such as particle size, shape, and texture, influence the performance of the final product.Conclusion:In conclusion, civil engineering materials are diverse and essential for the construction industry. Concrete, steel, brick, timber, asphalt, and aggregates all have unique properties that make them suitable for different applications. Understanding the characteristics of these materials is crucial for designing and constructing safe and durable structures. By utilizing the right materials and considering factors such asstrength, durability, and cost-effectiveness, civil engineers can ensure the success of their projects.。
形状英文的作文三年级简短Shapes are all around us, from the natural world to the man-made structures we create. They are a fundamental part of our everyday lives, and understanding and recognizing different shapes is an important skill for young learners. In this essay, we will explore the significance of shapes, their various classifications, and their practical applications in the world.Shapes are the foundation of our visual perception. They help us make sense of the world by breaking down the objects and structures we see into recognizable forms. Whether it's a circle, a square, a triangle, or a more complex geometric shape, our ability to identify and categorize shapes is crucial for understanding the environment around us.One of the most basic classifications of shapes is the distinction between two-dimensional (2D) and three-dimensional (3D) shapes. Two-dimensional shapes, such as circles, rectangles, and triangles, exist on a flat plane and have length and width but no depth. In contrast, three-dimensional shapes, like spheres, cubes, andpyramids, occupy space and have length, width, and depth.Understanding the properties of different shapes is essential for a wide range of practical applications. In the field of architecture, for example, architects and engineers rely on their knowledge of shapes to design sturdy and visually appealing structures. The use of triangular, rectangular, and circular shapes in building foundations, walls, and roofs contributes to the overall stability and aesthetic of a structure.Similarly, in the realm of art and design, shape is a fundamental element that artists and designers use to create visually compelling compositions. Shapes can be used to convey emotions, evoke specific moods, or simply to organize the elements of a design in a harmonious and aesthetically pleasing way.Beyond the realms of architecture and art, shapes also play a crucial role in various scientific and technological fields. In mathematics, the study of shapes and their properties, known as geometry, is a core component of the curriculum. Students learn to identify, classify, and analyze different shapes, as well as to understand the relationships between them.In the field of engineering, the study of shapes is essential for the design and construction of a wide range of products, from everydayhousehold items to complex machinery and vehicles. Engineers must consider the shape and structure of the materials they use, as well as the overall shape of the final product, to ensure it is functional, durable, and efficient.The importance of shapes is not limited to the physical world; it also extends to the digital realm. In computer graphics and digital design, the manipulation and rendering of shapes are fundamental to the creation of visual content, from simple icons and logos to highly detailed 3D models and animations.Furthermore, the recognition and understanding of shapes are crucial in the development of artificial intelligence and machine learning algorithms. These technologies rely on the ability to identify and classify shapes in order to process and interpret visual information, which has a wide range of applications, from facial recognition to autonomous vehicle navigation.In conclusion, shapes are a fundamental aspect of our world, and understanding their properties and applications is essential for a wide range of fields and disciplines. From architecture and art to science and technology, the study of shapes has proven to be a valuable and enduring area of exploration and discovery. As we continue to explore and push the boundaries of our understandingof shapes, we can expect to see even more innovative and exciting applications emerge in the years to come.。
Understanding the Shape Properties of Trihedral PolyhedraCharlie Rothwell and Julien SternINRIA,2004,Route des Lucioles,Sophia Antipolis,06902CEDEX,FranceAbstract.This paper presents a general framework for the computation of pro-jective invariants of arbitrary degree of freedom(dof)trihedral polyhedra.Weshow that high dof.figures can be broken down into sets of connected four dof.polyhedra,for which known invariants exist.Although the more general shapesdo not possess projective properties as a whole(when viewed by a single cam-era),each subpart does yield a projective description which is based on the but-terfly invariant.Furthermore,planar projective invariants can be measured whichlink together the subparts,and so we can develop a local-global description forgeneral trihedral polyhedra.We demonstrate the recovery of polyhedral shape de-scriptions from images by exploiting the local-global nature of the invariants.1IntroductionIn this article we introduce a general scheme for understanding the shape properties of trihedral polyhedra.Trihedral polyhedra are solid polyhedra made up of planes in arbi-trary positions,and as such,no special constraints exist between the planes.The nomen-clature trihedral derives from the fact that the vertices of the polyhedra are only ever defined by triples of planes:points in space need at least three planes to assert their lo-cations,but any more would provide excess constraint and hence would not be generic (and stably realisable).The results in this paper are a summary of those given in[9].In all,we generalise the result in[8]which showed how a projectively invariant de-scription can be computed for four degree of freedom(dof)polyhedra from a single view.In turn,[8]was a extension of the work of Sugihara[11].The latter dealt with scaled orthographic projection and the calibrated perspective cases,whereas the former demonstrated the projective equivalence of all members of the family of four dof.poly-hedra generating a set of scene measurements using an uncalibrated camera.We show in this paper that the approach of[8]can be extended to include all trihedral polyhedra.We also build on some recent work for computing the invariants of minimal point configurations in three-dimensional space.Being able to compute measures for small local feature groups provides robustness to occlusion.More global descriptions can be built up using the local-global nature of many shape descriptions[3,10].We derive a part-whole decomposition by drawing the invariant description of[8],and the invari-ants based on the butterfly configuration of Mundy[13]together.The butterfly invariant is a geometric description of a special six-point configuration.Our interest in the butterfly invariant was promoted by the recent paper of Sugi-moto[12].This paper discusses an invariant very similar to the original butterfly in-variant,but suggests an algebraic rather than a geometric formulation.However,Sugi-moto suggested that the invariants in[12]in some way replace the invariants described by[8].In fact,these two types of invariant can be taken hand-in-hand and are exactly complementary.This we show partly in this paper,and in more detail in[9].The contributions of this paper are three-fold:in Section2we discuss how the orig-inal invariant description of[8]can be decomposed into a set of three independent but-terfly invariants.Then we show in Section3how to reduce afive dof.figure into sets ofA B E FQPDCl l l 312A GDB EC F (a)(b)Fig.1.(a)the butterfly consists of two sets of four coplanar points yielding a unique cross ratio.(b)the labelling of a cube used for the butterfly invariant computation.figures with four dof.For the simplest case of such an object (two cube-like structures stuck together using a common plane),we can recover three invariants for each half of the object,and a further eight joint-invariants based on planar measures between the two halves.We also describe how higher dof.polyhedra can be broken down into four dof.objects related through pairwise planar constraints.Finally,in Section 4we report on algorithms for the extraction of the polyhedral descriptions and their image invariants.2The butterfly invariantThe butterfly is the simplest known invariant for a set of points in 3D space.1The con-figuration is composed of six points in space broken up into two four point groups,and .Each four point group is coplanar,and two points are shared between the groups.This is shown in Fig.1a.The invariant for the butterfly is measured through the construction of a cross ratio.As can be seen in Fig.1a,it is pos-sible to form a set of four collinear points and hence the invariant cross ratio [2].These are the points from which we define the cross ratio .In fact,we show in [9]that the same butterfly invariant actually takes the algebraic form:(1)which allows the direct computation of the invariant values from image data.Hereis the matrix whose columns are the points ,and and is the determinant of M .Although this invariant is very similar to that derived by Sugimoto [12]the interested reader will note that it can be derived more simply (shown in [9]).It is also worth noting that the form above was derived by Carlsson [1]using a more difficult,but in fact more general (and so elegant)approach based on the double algebra.2.1Computing polyhedral invariants with the butterflyThe projective invariants for polyhedral figures can be computed using the butterfly.Here we consider how to compute the invariants for a cube-like structure such as that shown in Fig.1b.In fact,we see later that understanding the invariants for this type of figure is as far as we need to go to comprehend the invariants for all trihedral polyhedra.For a cube-like polyhedron,there are three independent invariants,for instance:(2)It is simple to demonstrate that these three invariants are independent and form a basis for all of the other invariants.Thus a description based on the butterfly is equivalent to one which uses the invariants of[8].Proofs of these statements are given in[9].3Simplifying higher order polyhedraWe now extend previous theories on trihedral polyhedra to the general case.This result is significant because it resolves a number of important but unanswered questions which were posed by[8].There,it was shown that invariants can be computed for polyhedral figures which have four dof.These represent only a relatively simple class of polyhedra such as cubes,as well as shapes equivalent to cubes but with simple volumes cut-out of them.Note that the number of dof.of a polyhedra represent the dimensionality of the space of reconstructions which arise out of the image constraints.In the following paragraphs we demonstrate thatfigures with more than four dof. can be decomposed into sets of related polyhedra which have only four dof.Thefirst example we consider is the decomposition offive dof.figures into sets of four dof.fig-ures.More complicatedfigures can be decomposedsimilarly.The simplification process means that the invariants we have seen previously for four dof.polyhedra(such as the butterfly invariants)can be employed for each subpart of a more complicated structure, and so informs us that the basic invariants for general polyhedral structures are based on those for four dof.figures.However,no projective invariants of the complete structures exist as each four dof.subpart maintains a certain amount of(non-projective)indepen-dence from the other parts.We discuss this claim in Section3.3.Additionally,we can compute joint-invariants between the pairs of adjacent four dof.figures in the subpart hierarchy;these joint-invariants provide the glue which holds the subparts together.Therefore,although there are no general projective constraints between the subparts of a polyhedron,the subparts are not reconstructed in space with total liberty,but are placed in related ually,each related pair of subparts in a figure share a common plane in which plane projective invariants enforce constraints.These two sets of invariants are all those which may be computed for an arbitrary trihedral polyhedron.We now proceed with proofs and explications of the statements given above.Due to limitations of space we are unable to repeat the theory given in[8], but only highlight the key points.However,we do make use of the same notation. 3.1Reconstructing polyhedralfiguresThe assumptions we make about the polyhedra we treat are that they are made up of trihedral vertices and thus satisfy genericity conditions.Additionally,they are bounded solids which,due to self-occlusion under imaging,can be considered not to have interior volumes or surfaces.Thus,any complete set of adjacent planes forms a solid polyhe-dron.A complete polyhedron has three planes meeting at every vertex,and two planes at every edge.We can always render a set of adjacent planes complete using the trihedral constraint by defining additional planes passing through the edges and vertices.We use a3D coordinate system and so planes satisfy:,where is the number of planes on the object.Each of the represent the polyhedral vertices.Under the pinhole projection model, projection onto the plane maps the point to where: and.Substituting these into the plane constraint equation,dividing through by and setting yields:(3)Here we have assumed a camera calibration;as stated in[8],this need not be the case. The trihedral assumption means that each point lies on three planes,say. Eliminating the between pairs of constraints yields two linear equations per point:(4) Now,from the image vertices observed in each image,we form which is a matrix equation of constraints in unknowns.Here, ,is a matrix,and.We now state the following theorem:Theorem3.1The kernel of the matrix A represents all of the possible solutions to the reconstruction of a trihedral polyhedron imaged by a single pinhole camera.Proof See[8]and more completely[7].The kernel of A can be represented as.We can orient the basis for w so that for and where,and.We may alsofix the other basis vectors so that:for.This means that the solution for w with for,represents a simple planar configuration(strictly at least one of should satisfy for). The initial three elements of being zero means that thefirst plane of the reconstructed polyhedron has coordinates.3.2Four dof.polyhedronIt was shown in[7]that for the solutions represented by w are all projectively equivalent to each other.In this case we thus need consider only one solution from w for the computation of the three-dimensional projective invariants on the shape.3.3Five dof.polyhedronConsider the case for.To simplify the discussion we represent and by:andSeparating out thefirst four dof.figureWe return to the form of the constraints given in eqn(4).Consider thefirst plane,and another plane which shares an edge with it;the endpoints of the edge are the points and.First we construct a set of planes which initially contains thefirst and the planes.Consider writing.Then,making use of the fact that both and are null vectors means that and. Subsequently,eqn(4)reduces to the following vectorial constraints:(5) These define uniquely up to a scale as and are distinct points.This solution exists for all and,and so either is parallel to,or one of and is null(but not both).Now,within the bounds of the geometric constraints used so far,weare free to orient and as we desire.In order to simplify the overall description, we make a further change of basis.Assuming that and are parallel,a change of basis is effected so that becomes a null vector,and is non-zero.Should either of the two vectors have already been null we can make a change of basis(perhaps trivially) to ensure that is non-zero and is null.Once done,neither thefirst plane,nor plane, will depend on.By reapplying the reasoning,we may extend the process and provide independence to for all of the other planes which have parallel to(or are null vectors).All planes of this type are included in the set.It is certain that a number of planes will be included in.Geometrically we see that providing independence to for other planes is based on the trihedral assumption underlying our manipulations.Consider any third plane which shares an edge with both thefirst and planes;the third plane is precisely defined by the positions of the other two.Thefirst plane is independent of both and,and the one depends only on .Thus the new plane depends only on(and so its s must be null).This process can be continued for all other planes which have two edges in common with a pair of planes from.One will thus build up a set of planes constituting which depend only on ,and which consequently represent a four dof.figure.The second four dof.figureOnce thefirst four dof.subpart has been extracted we can be sure that no planes remain outside which are in contact with more than one plane in.This is because if a plane not in were attached to two planes in,then it would have to befixed in space.We label the complementary set to by.We may now make the following statements: 1.The planes in use up four of thefive dof.for the polyhedron.This is because theplanes in are defined by the,.2.does not represent the entire polyhedra(is not empty).If it did,then wouldbe null and we would have only a four dof.figure.3.There must be at least one plane in which is adjacent to a plane in.If this werenot the case,then the planes in would make up afigure with at least three more dof.These extra three dof.would lead to the entirefigure having too many dof. We exploit the third of these statements to base the rest of the reconstruction around the plane in which is adjacent to.To do this,we again change basis so that this plane is accounted for by thefirst three parameters of w.For simplicity of argument we also reorder the rest of planes so that those dependent on only appearfirst in the basis vectors.The plane(which is thefirst plane that does not belong to)is chosen so that is shares an edge with thefirst plane.Thus we have:We now repeat the process used previously for the elimination of the components of from for,though this time we eliminate the.Given that thefirst plane is rep-resented by the zero vector,and that the plane shares an edge with it,we know that and are parallel(cf.eqn(5)).We may now eliminate either of or to yield dependence only on one of or.Clearly this plane cannot depend on otherwise it would already have been included within.Therefore,we are able to eliminate. In the same way that we added planes to we continue to build up planes dependent only on.We place all of these planes in the set.Once all possible have been setto zero,we are left with basis vectors of the form:(6) Such a basis represents a decomposition of the polyhedron into two four plete solid polyhedra built around thefirst plane(thefirst polyhedron of planes1to, and the second of plane1,and planes to).Thefirst plane is called the common plane.In addition to these sub-figures,there may be a set of planes to which we represent by the set.is empty for simple polyhedra,and so.However, this is not the general case and can well contain further polyhedral subparts.Note that the planes in this subsidiary set are still constrained completely by and.atfirst has the appearance of being a more complexfigure than the four dof.sub-parts which we have already extracted,though this is not really the case.will always be composed of other four dof.polyhedra.The simplest way to understand this is to re-order the planes so that one of the planes in adjacent to is set to thefirst plane of w(such a plane must exist).We then repeat the above elimination process.Doing this would mean that a subset of planes in would be dependent only on the new(is still parametrized by)and so would have only four dof.By progressive choices of common planes between the extracted subparts of and their complement in,we can gradually demonstrate that is composed only of four dof.figures.The trick is to see that any two polyhedra sharing a common plane can always be demonstrated to have four dof.We thus extend our current understanding of trihedral polyhedra via: Theorem3.2Five dof.trihedral polyhedra are made up of simple building blocks con-sisting only of constrained four dof.polyhedra.As each subpart has only four dof.,their own local equivalence classes are projective.Proof As discussed above,and more completely in[9].Projective inequivalence offive dof.figuresUnlike the four dof.figures discussed previously in[8],two different reconstructions of afive dof.figure need not necessarily be projectively equivalent.Formally: Theorem3.3The different reconstructions of afive dof.non-trivial polyhedron imaged by an uncalibrated pinhole camera are not in the same projective equivalence class.Proof Given in[9].3.4Invariants forfive dof.figuresThere are two sets of invariants which can be measured forfive dof.polyhedralfigures. Thefirst are based on the invariants of each of the subparts,and so are the same as the four dof.invariants given in[8].As stated in Section2.1,these can be measured using the butterfly invariants.We can also measure a number of invariants in the common plane from the coplanar edges of each adjacent sub-polyhedra.Generically the number of invariants of this form depends on the structure of the polyhedra:if there are edges in each of the common planes between the subparts,,then there are computable planar invariants.It can be shown that the butterfly and planar invariants represent all of the invariants which can be computed for such afigure.3.5Higher degree of freedomfiguresWe can use the results of Section3.3to complete the description paradigm for polyhedra of six or more dof.We proceed with higher dof.figures in a similar way tofive dof. polyhedra.First,extract a four dof.subpart and define the planes which it has in common with the rest of the polyhedron.Then,continue to examine the rest of the polyhedron by ignoring the planes in thefirst subpart,extracting further four dof.polyhedra.In this manner the whole polyhedron can be decomposed into a set of four dof.subparts.The computation of the invariants for the entire object is then straightforward:–Compute projective invariants for the planes within each subpart.–Compute planar invariants between the edges in all the common planes.It follows obviously from the result forfive dof.polyhedra that the families of recon-structions for higher dof.figures are not projectively equivalent.4Finding polyhedra in imagesIn this section we report on a pair of approaches which show how polyhedral descrip-tions can be extracted from real images.As always,we are plagued by the difficulty of extracting accurate segmentations from real images.We discuss two different methods and provide demonstrations of them working on relatively simple images containing well-defined polyhedral objects.The solutions are:–Use an edgel detector to initialize a polyhedral snake on sets of image features.Then compute invariants for the snake.–Search for pairs of adjacent closed regions in the image suitable for estimating butter-fly invariants.These invariants are used to index into a model base and subsequently to provide hypotheses suggesting which object might be present in the scene.Hy-potheses are combined post-indexing to derive richer object descriptions.Both of these approaches require edgel and line extraction.The edgel detector we use is described in[6]and ensures good connectivity around image junctions.Straight lines arefitted to edgel-chains using orthogonal regression and a topological(connectivity) structure composed of connected straight lines and edgel-chain segments is produced. The actual data structures we use are built around a vertex-edge-face topology hierar-chy[5].Vertices are typically used to represent junctions and the interfaces between pairs of lines or edgel-chains.Edges link vertices and are instantiated geometrically by lines and edgel-chains.Faces represent closed cycles of edges(we also have1-chains which represent non-cyclic edge chains).This topology hierarchy allows the straight-forward extraction of closed regions from the image as each face is a closed region. 4.1Polyhedral snakesThe primary goal of the snake extraction approach is to recover structures which have the same topological forms as the projections of the polyhedra which we wish to recog-nize.Once these have been extracted,they can be matched to snake models,and then the snakes develop in the image by interaction with the image intensity surface.After a number of iterations we can measure the invariants of the snake(which are similar to projected polyhedra),and thus hypothesize the identity of the object in the scene.Thefirst phase of processing involves the extraction of edges,as demonstrated in Fig.2.Notice that we have been able to recover a fairly good description of the polyhe-dron’s topology with the edge detector.Principally the edge description consists of three closed regions each of which match projected polyhedral faces.Being able tofind these image regions means that we can reject scene clutter(non-polyhedral regions)rapidlyFig.2.First we show a polyhedron and then the output of the edgel detector superimposed.Note that the basic topological structure of the edges is correct.Next are the linesfitted by orthogonal regression andfinally the polyhedral snake.1C C2C3C4C5C6C7 Fig.3.The topological descriptions stored for the projections of a six-plane polyhedron. during subsequent processing.We thenfit lines and throughout ensure the maintenance of the topological description.Afterfitting,we extract faces from the image.For Fig.2 there are three basic regions which are suitable,plus a number of others which include combinations of these regions.Subsequent analysis is focused on the faces and on the lines which they contain.However,over-segmentation can arise due to the edge detec-tor recovering small noisy features.A number of these are shown in Fig.2.These fea-tures disrupt the linefitting process and cause breaks to be inserted between pairs of line segments.We therefore traverse the boundary of each face and test whether a pair of adjacent lines would be better represented by a single straight line segment.If so,a sin-gle line segment is substituted.Once the line merging process is complete,we count the number of lines in each face to see whether it remains of interest.Polyhedral faces must contain at least three line segments,and so any with less can be discarded.For this appli-cation we desire at least four lines.Consecutive line segment endpoints should also be reasonably close as polyhedral faces are generally polygonal.Given this,we can discard the edgel-chains between lines and reduce the face representations to ideal polygons.The penultimate stage of processing involves the matching of the adjacency graph of the faces to topological models which we have for the polyhedra in the model base.The topological descriptions of the lines and the snake are consistent in Fig.2.The complete set of descriptions from which we form snakes is given in Fig.3.Given a match,such as that shown,we initialise a snake which is allowed to relax onto the contrast boundaries in the original image.The snakes are a crude implementation of those described in[4]. Thefinal position of the snake yields sufficient structure for invariant computation.In Fig.2,all of the butterfly invariants are unity because the polyhedron is projectively equivalent to a cube(which is composed of three of sets of parallel planes).Table1 shows how the three invariants for the shape change over the iterations of the snake.A similar example of polyhedron extraction using snakes is in Fig.4.In Fig.5we have a harder case.The edgel detector failed to recover the polyhedron’s vertical inter-nal boundary,and so we must use a simplified model with which to initialize the snake. In this case we have a description equivalent to(in Fig.3)and so we are able to hypothesize the location of the missing edge by joining up a pair of vertices.We re-turn to the complete snake description()once the missing edge has been inserted.Initeration invariant20-0.897423-0.956112-1.04808 8-1.01821Fig.6.After segmentation and closed region detection we extract three butterflies with invariants approximately equal to unity.These butterflies can be merged into a single polyhedral description.Fig.7.We recover three butterflies for this imaged polyhedron.principal we can re-apply the hypothesis extension process between the four dof.figures by using the planar projective invariants in the common planes as tests of consistency.Extraction of the butterflies is shown in Fig.6.We show the three different extracted butterflies for the image section in thefigure.The butterflies are composed of three im-age faces each containing four principal line segments,and have invariant values of 1.007,1.022,and1.000respectively.All of these match butterfly invariants for objects which have projective equivalence to a cube.They can thus be merged into the single polyhedral description shown on the right which is a single four dof.polyhedron.Another example is in Fig.7.There the three butterflies have invariants equal to 1.000,1.000,and1.029.However,in Fig.8we show an example in which only a single butterfly can be recovered(with an invariant of1.000)due to the failure of the edgel detector to recover more than two of the polyhedral faces as closed regions.In this case we have only weak support for a polyhedral hypothesis as one matching invariant carries less weight than three.Nevertheless,the precision of the single invariant suggests that the features probably do correspond to a polyhedron with similarity to a cube.5ConclusionsWe have made a number of different contributions in this paper.Thefirst two are at a theoretical level,and the third expresses a pair of practical approaches for the extrac-tion of polyhedral descriptions from images.Thefirst theoretical study presented a veryFig.8.For this polyhedron we recover only two polyhedral faces as closed regions in the edgel description.Thus we canfind only a single butterfly configuration.However,the invariant value for the butterfly matches precisely that which we would expect tofind for such a polyhedron.simple algebraic formulation for the butterfly invariant and showed how it is related to the polyhedral invariants of[8].The second theoretical contribution is a completion of the projective invariant de-scription of[8]to trihedral polyhedra of arbitrary degrees of freedom.We demonstrated that all higher order polyhedra can be broken down into sets of four dof.polyhedra as-sociated using plane projective invariants.This yields a complete description paradigm for arbitrary trihedral polyhedra.Finally,we developed two different implementations of algorithms intended for the detection of polyhedra in images.Thefirst is founded on topological reasoning which is used as a basis for a polyhedral snake process.The second makes use of closed region detection for the creation of local hypotheses,and then the use of hypothesis extension for the generation of more complete polyhedral descriptions.A more complete version of this paper[9]is available as a technical report from the URL:<URL ftp://ftp.inria.fr/INRIA/publication/publi-ps-gz/RR/RR-2661.ps.gz>. AcknowledgmentsWe acknowledge the input of a number of people in this work,principally:Olivier Faugeras, David Forsyth,Bill Hoffman,Joe Mundy,and Andrew Zisserman.Thanks to Pippa Hook for proof-reading this article.CR is funded by a Human Capital and Mobility grant from the Euro-pean Community.JS was visiting INRIA from the Ecole Normale Supe´rieure de Lyon. References[1]S.Carlsson.Multiple image invariance using the double algebra.In Applications of Invari-ance in Computer Vision,volume825of LNCS,p.145–164.Springer-V erlag,1994.[2]R.Duda and P.Hart.Pattern Classification and Scene Analysis.Wiley,1973.[3]rge hierarchical object recognition using libraries of parameterized modelsub-parts.Proc.CVPR,p.32–41,1988.[4]M.Kass,A.Witkin,and D.Terzopoulos.Snakes:Active contour models.Proc.ICCV,p.259–268,1987.[5] C.Rothwell,J.Mundy,and W.Hoffman.Representing objects using topology.In prepara-tion,1996.[6] C.Rothwell,J.Mundy,W.Hoffman,and V.-D.Nguyen.Driving vision by topology.Proc.IEEE International Symposium on Computer Vision,p.395–400,1995.[7] C.Rothwell,D.Forsyth,A.Zisserman,and J.Mundy.Extracting projective informationfrom single views of3D point sets.TR1927/92,Oxford Univ.Dept Eng.Sci.,1992. [8] C.Rothwell,D.Forsyth,A.Zisserman,and J.Mundy.Extracting projective structure fromsingle perspective views of3D point sets.Proc.ICCV,p.573–5821993.[9] C.Rothwell and J.Stern.Understanding the shape properties of trihedral polyhedra.TR2661,INRIA,1995.[10] C.Rothwell.Object recognition through invariant indexing.Oxford University Press,1995.[11]K.Sugihara.Machine Interpretation of Line Drawings.MIT Press,1986.[12] A.Sugimoto.Geometric invariant of noncoplanar lines in a single view.Proc.ICPR,p.190–195,1994.[13] A.Zisserman,D.Forsyth,J.Mundy,C.Rothwell,J.Liu,and N.Pillow.3D object recog-nition using invariance.TR2027/94,Oxford Univ.Dept Eng.Sci.,1994,to appear the AI Journal.This article was processed using the L A T E X macro package with ECCV’96style。