Resolution-Based Approximate Reasoning
for OWL DL
Pascal Hitzler and Denny Vrandeˇc i′c
AIFB,Universit¨a t Karlsruhe,Germany
{hitzler,vrandecic}@aifb.uni-karlsruhe.de
Abstract.We propose a new technique for approximate ABox rea-
soning with OWL DL ontologies.Essentially,we obtain substantially
improved reasoning performance by disregarding non-Horn features of
OWL DL.Our approach comes as a side-product of recent research
results concerning a new transformation of OWL DL ontologies into
negation-free disjunctive datalog[1,2,3,4],and rests on the idea of per-
forming standard resolution over disjunctive rules by treating them as
if they were non-disjunctive ones.We analyse our reasoning approach
by means of non-monotonic reasoning techniques,and present an imple-
mentation,called Screech.
1Introduction
Knowledge representation and reasoning on the Semantic Web is done by means of ontologies.While the quest for suitable ontology languages is still ongoing, OWL[5]has been established as a core standard.It comes in three?avours, as OWL Full,OWL DL and OWL Lite,where OWL Full contains OWL DL, which in turn contains OWL Lite.The latter two coincide semantically with certain description logics[6]and can thus be considered fragments of?rst-order predicate logic.
OWL ontologies can be understood to consist of two parts,one intensional, the other extensional.In description logics terminology,the intensional part con-sists of a TBox and an RBox,and contains knowledge about concepts(called classes)and the complex relations between them(called roles).The extensional part consists of an ABox,and contains knowledge about entities and how they relate to the classes and roles from the intensional part.For the Semantic Web, TBox and RBox shall provide background vocabulary,while(annotated)web-pages etc.constitute ABoxes which are interlinked with intensional knowledge. The Semantic Web thus envisions a distributed knowledge source,built from OWL ontologies and intertwining the knowledge like the World Wide Web in-terconnects websites.
The authors acknowledge support by the German Federal Ministry of Education and Research(BMBF)under the SmartWeb project,and by the European Commission under contract IST-2003-506826SEKT and under the KnowledgeWeb Network of Excellence.The expressed content is the view of the authors but not necessarily the view of any of the projects as a whole.
Y.Gil et al.(Eds.):ISWC2005,LNCS3729,pp.383–397,2005.
c Springer-Verlag Berlin Heidelberg2005
384P.Hitzler and D.Vrandeˇc i′c
With an estimated25million active websites today and correspondingly more webpages,it is apparent that reasoning on the Semantic Web will have to deal with very large https://www.doczj.com/doc/2a4117938.html,plexity of ABox reasoning—also called data com-plexity—thus measures complexity in terms of ABox size only,while considering the intensional part of the ontology to be of constant size.For the di?erent OWL variants,data complexity is at least NP-hard,which indicates that it will not scale well in general[7].Methods are therefore being sought to cope with large ABoxes in an approximate manner.
The approach which we propose is based on the fact that data complex-ity is polynomial for non-disjunctive datalog.We utilise recent research results [1,2,3,4]which allow the transformation of OWL DL ontologies into disjunc-tive datalog.Rather than doing(expensive)exact reasoning over the resulting disjunctive datalog knowledge base,we do approximate reasoning by treating disjunctive rules as if they were non-disjunctive ones.The resulting reasoning procedure is complete,but may be unsound in cases.Its data complexity is polynomial.We are also able to give a characterization of the resulting ap-proximate inference by means of standard methods from logic programming semantics.
This paper is structured as follows.In Section2,we?rst discuss the general rationale behind approximate reasoning,and how it relates to other reasoning frameworks.We then recall formal terminology and notation for OWL DL,and shortly review datalog and SLD-resolution.Then,in Section4,we explain how OWL DL ontologies can be transformed into disjunctive datalog.In Section5 we introduce the new approximate SLD-resolution procedure which we propose. The presentation of our implementation Screech in Section6is followed by an Example in Section7,and an experimental evaluation in Section8.We conclude and discuss future work in Section9.
2Non-classical Reasoning—Common Grounds
The sophisticated reasoning tasks required when dealing with expressive knowl-edge representation languages like those based on description logics are known to be of high computational complexity.In the face of ever increasing data quan-tities to be processed,new methods are needed to obtain usable systems.As the high computational complexity of the reasoning tasks is unavoidable,the method of choice for obtaining scalable systems is to use approximate reasoning techniques.In a nutshell,approximate reasoning rests on the idea of decreas-ing the complexity of a problem by imposing controlled changes on either the language used or the inference operation used for the deduction.The resulting lower complexity and consequent speed-up thus comes at the price of unsound-ness or incompleteness(or both),but in a controlled and well-understood manner which allows to assess the quality of the deduction made by the approximate rea-soner.So-called anytime algorithms develop the idea a bit further and guarantee convergence to exact answers given enough time,while providing approximate results during the reasoning process.
Resolution-Based Approximate Reasoning for OWL DL385
Fig.1.Semantic view on approximate reasoning
A semantic perspective on approximate reasoning is depicted in Figure1. When a theory is being considered,classical reasoning may be of high com-putational complexity and thus be unsuitable for time-critical tasks.By taking di?erent models into account than the classical ones,the complexity of reasoning can be reduced.The resulting approximate inference may be incomplete or un-sound with respect to classical inference,but in a controlled and well-understood manner,which makes the inferences suitable for further use.
Similar situations occur in the context of other sophisticated reasoning tech-niques.For non-monotonic reasoning,for example,a subset of the classical mod-els is usually considered,which is selected by means of e.g.additional syntax constructs or by rede?ning the semantics of existing ones.Non-monotonic rea-soning thus allows to arrive at conclusions which cannot be derived using classical reasoning:It is complete,but unsound,and can be described as supraclassical[8]. The rationale in this case is to model aspects of human commonsense reasoning like jumping to conclusions,again in a controlled and well-understood manner. Complexity considerations are often treated as secondary in this context.
Paraconsistent reasoning—or reasoning with inconsistency—can be ap-proached from a similar perspective.While inconsistent knowledge bases have no classical models,paraconsistent reasoning strives to identify suitable models to be assigned to the knowledge base nevertheless,in order to allow the infer-ence of meaningful consequences.As such,paraconsistent reasoning is sound,but incomplete with respect to classical logic,and can thus be termed subclassical.
Table1summarizes our discussion.While the table can certainly be extended further taking other forms of reasoning into account,we restrict ourselves to the mentioned examples,as the main goal of this paper is to present an approxi-mate reasoning method for OWL DL,and not a comparative theory of reasoning https://www.doczj.com/doc/2a4117938.html,parision of non-classical reasoning approaches
reasoning approach focus models taken into account typical complexity classical all classical models high
non-monotonic commonsense some classical models very high paraconsistent inconsistency more than the classical models high
approximate performance variable low
386P.Hitzler and D.Vrandeˇc i′c
approaches.We have included this discussion because it explains the general ra-tionale behind our approximate reasoning method,and will help us in analyzing it.Indeed,in all reasoning paradigms mentioned,it is important to obtain a clear understanding of the inference relation computed.This can be done by semantic analyses,i.e.by characterizations of the models taken into account.From the general perspective described in this section,it will later come as no surprise to the reader that we will analyze our approximate reasoning methods by means of standard techniques from non-monotonic reasoning.Indeed,in our particular case the models taken into account for approximate reasoning will turn out to be a subset of the classical models,as in non-monotonic reasoning.
3Preliminaries
3.1OWL DL Syntax and Semantics
OWL DL is a syntactic variant of the SHOIN(D)description logic[9].Hence, although several XML and RDF syntaxes for OWL DL exist,it will be convenient to use the traditional description logic notation since it is more compact,and we recall the notation below.For the correspondence between this notation and various OWL DL syntaxes,see[9].
We indeed assume that the reader is familiar with OWL and thus with SHOIN(D),as space restrictions forbid to reintroduce them,but recall that SHOIN(D)supports reasoning with concrete datatypes,such as strings or inte-
gers[10].Recall also that the description logic syntax for concepts in SHOIN(D) is de?ned as follows,where A is an atomic concept,R is an abstract role,S is an abstract simple role,T(i)are concrete roles,d is a concrete domain predi-cate,a i and c i are abstract and concrete individuals,respectively,and n is a non-negative integer:
C→A|?C|C1 C2|C1 C2|?R.C|?R.C|≥n S|≤n S|{a1,...,a n}| |≥n T|≤n T|?T1,...,T n.D|?T1,...,T n.D
D→d|{c1,...,c n}
The SHIQ(D)description logic is obtained from SHOIN(D)by disallowing nominal concepts of the form{a1,...,a n}and{c1,...,c n},and by allowing qual-i?ed number restrictions of the form≥n S.C and≤n S.C,for C a SHIQ(D) concept and S a simple role.
As description logics,SHOIN(D),i.e.OWL DL,and SHIQ(D)inherit their semantics from?rst-order logic by the standard translations known e.g. from[11],which we do not repeat here.
3.2Datalog and SLD-Resolution
A(de?nite or negation-free)disjunctive logic program P consists of a?nite set of clauses or rules of the form
?x1...?x n.(H1∨···∨H m←A1∧···∧A k),
Resolution-Based Approximate Reasoning for OWL DL387 commonly written as
H1∨···∨H m←A1,...,A k,
where x1,...,x n are exactly all variables occuring in H1∨···∨H m←A1∧···∧A k,and all H i and A j are atoms over some given?rst-order language Σ.The disjunction H1∨···∨H m is called the rule head,and the conjunction A1∧···∧A k is called the rule body.The set of all ground instances of atoms de?ned overΣis called the Herbrand base of P and is denoted by B P.The set of all ground instances of rules in P is denoted by ground(P).A rule is said to be non-disjunctive if m=1.It is called a fact if k=0.We abstract from the order of the atoms in the heads respectively bodies;it is not important for our results.A disjunctive logic program is called a(disjunctive)datalog program if it does not contain function symbols.
Note that we do not consider logic programs to come with one speci?c se-mantics.Some people for example associate datalog with the minimal model semantics only.For our treatment,datalog and logic programs are de?ned via syntax only.We do not specify a speci?c semantics because in the following we will discuss di?erent semantics for logic programs in their relation to proof procedures.One of the semantics we will consider is the semantics coming from interpreting logic programs as a set of?rst order formulas,and in this case we use|=to denote entailment in classical?rst-order predicate logic.
SLD-resolution(see e.g.[12])is an e?cient top-down query-answering tech-nique for programs consisting of non-disjunctive rules,and has been imple-mented and successfully applied in standard Prolog systems.1In this framework, a ground atom can be derived from a program if and only if it is true in the least (and thus in all)Herbrand models of the program.
In the following,we mean by a conjunctive query simply a conjunction B1∧···∧B n of atoms.The query is called ground if it does not contain any variables.
Given a conjunctive query B1∧···∧B n,an SLD-resolution step on the atom B i with a non-disjunctive rule H←A1,...,A k produces a conjunctive query
B1θ∧···∧B i?1θ∧A1θ∧···∧A kθ∧B i+1θ∧···∧B nθ
whereθis the most general uni?er of B i and H.An SLD-refutation of a con-junctive query B1∧···∧B n in a non-disjunctive program P is a?nite sequence of conjunctive queries Q0,...,Q n,where(i)Q0=B1∧···∧B n,(ii)each Q i with i>0is obtained from Q i?1by an SLD-resolution step with some rule from P on some literal B i,and(iii)Q n= ,i.e.the conjunctive query Q n does not contain any literals.If an SLD-refutation of B1∧···∧B n in P exists,we write P B1∧···∧B n.
One of the fundamental results in logic programming states that A∈B P can be proven by SLD-resolution if and only if A is a logical consequence of P,i.e. if and only if A is true in the least Herbrand model of P:
1Like SWI or XSB Prolog,https://www.doczj.com/doc/2a4117938.html,,https://www.doczj.com/doc/2a4117938.html,.
388P.Hitzler and D.Vrandeˇc i′c
Theorem1([12]).For a ground conjunctive query B1∧···∧B n and a non-disjunctive program P,P B1∧···∧B n if and only if P|=B1∧···∧B n.In other words,entailment of ground conjunctive queries under SLD-resolution is entailment in predicate logic.
SLD-resolution also allows to derive answers to non-ground queries:For a con-junctive(and not necessarily ground)query Q there exist an SLD-refutation if and only if P|=?x1...?x n.Q,where x1,...,x n are the variables occuring in Q.By keeping track of the most general uni?ers used in the process,it is also possible to obtain bindings for(some of)the x i in the form of(answer)substitu-tionsθ,such that P|=?y1...?y k(Qθ),where the y i are exactly those variables occurring in Qθ.In order to keep our exhibition focused,we will only deal with ground queries.
4Reducing OWL DL Knowledge Bases to Disjunctive Datalog Programs
We utilise recent research results about the transformation of OWL DL ontolo-gies into disjunctive datalog,and perform approximate reasoning by transform-ing the disjunctive database into a non-disjunctive one.The transformation is based on the fact that OWL DL is a subset of?rst-order logic.OWL axioms can thus be translated directly into logical formulas and transformed into clausal form using any of the standard algorithms.The resulting clauses can be repre-sented as disjunctive datalog rules which do not contain negation.
Note,however,that due to possible skolemization steps in the clausal form translation,the resulting datalog rules may contain function symbols.In general, datalog with function symbols is undecidable,but since we obtain the datalog program by a translation from OWL DL,which is decidable,inferencing over the resulting program must be decidable.Standard datalog engines,however,do in general not terminate in the presence of function symbols.To cope with this problem,a sophisticated method has been presented in[2,3]which allows to get rid of the function symbols without loosing ABox consequences.As a result,we obtain a function-and negation-free disjunctive datalog program,which can be dealt with using standard techniques.
There is one other catch:The approach presented in[2,3]does not yet allow to deal with nominals,i.e.it supports only SHIQ(D)instead of SHOIN(D) (the latter is the description logic coinciding with OWL DL).We remark that to date—and to the best of our knowledge—no reasoning algorithms for SHOIN(D)have been implemented.We will return to a possible treatment of nominals in our approach later.
The translation algorithm is schematically depicted in Figure2.It transforms a SHIQ(D)knowledge base KB into a disjunctive datalog program DD(KB). The steps of the algorithm are as follows.(1)Transitivity axioms are removed by adding axioms of a form similar to?S.C ?S.(?S.C)for transitive roles S.
(2)The knowledge base is translated into clausal form by standard transforma-tions based on?rst-order predicate logic.This introduces function symbols due
Resolution-Based Approximate Reasoning for OWL DL389
SHIQ(D)
KB
Disjunctive
Program
DD(KB) Fig.2.Algorithm for Reducing SHIQ(D)to Datalog Programs
to necessary skolemization steps.(3)The TBox of the knowledge base is par-tially saturated by adding logical consequences.This is the crucial step of the algorithm.(4)The saturation from step(3)now allows to remove all function symbols which were introduced in step(2).Some additional axioms are added to ensure that the algorithm remains sound and complete.(5)The knowledge base is translated into disjunctive datalog clauses;this step is now straightforward.
It shall be noted that the details of the crucial step(3)are very sophisticated. They guarantee that the removal of function symbols in step(4)is at all possible. Step(3)is of exponential complexity,however for the ABox reasoning task which we focus on in this paper,Step(3)can in principle be performed o?ine,as this step is independent of the ABox–but note that this o?ine computation may still be di?cult if the TBox is large,which is a seperate issue and deserves further in-depth studies which are outside the scope of this paper.A full presentation of the translation with correctness proofs is technically involved and lengthy, and space restrictions forbid to go into further detail;we refer the interested reader to[2,3].In[1]full proofs are given which show amongst other things that KB is unsatis?able if and only if DD(KB)is unsatis?able.This su?ces for reasoning over KB as reasoning tasks can be transformed into unsatis?ability checks.
5Approximate Resolution
While approximate reasoning methods for propositional and?rst-order logic have been proposed(see e.g.[13,14,15,16,17,18]),they have hardly been applied in the context of Semantic Web technologies.The few exceptions are reported e.g. in[19,20,21]—to the best of our knowledge,this list is exhaustive.The success of the approaches is mixed.[21]reports on an analysis indicating that straight-forward adaptations of methods proposed by[14]do not su?ce.[20]reports good results but is not an approximate reasoning method in the more narrow sense as the reasoning performed is exact,and thus does not address the complexity problems underlying OWL DL reasoning.[19]deals with approximating queries, while we focus on ABox reasoning.We will now present a novel approach based on the translation of OWL DL to disjunctive datalog,as presented earlier.
5.1Approximate SLD-Resolution
Having obtained the translated knowledge base in the form of a disjunctive datalog program,ABox reasoning remains NP-hard,and thus untractable.If the datalog program is non-disjunctive,though,reasoning is polynomial in the size of
390P.Hitzler and D.Vrandeˇc i′c
the ABox.We therefore propose the following approximate reasoning technique in order to facilitate this insight.Given a conjunctive query B1∧···∧B n, an approximate SLD-resolution step on the atom B i with a disjunctive rule H1∨···∨H m←A1,...,A k is a conjunctive query
B1θ∧···∧B i?1θ∧A1θ∧···∧A kθ∧B i+1θ∧···∧B nθ
such thatθis the most general uni?er of B i and some H j.Approximate SLD-refutation is de?ned analogously to SLD-refutation,where approximate SLD-resolution steps are used instead of(usual)SLD-resolution steps.
It is necessary to pursue the question what notion of entailment underlies the approximate reasoning technique we propose.Following the spirit of the obervations from Section2,we want to identify the set of models which underly the inference relation provided by approximate SLD-resolution.For this purpose, we need the following notion,which is derived from standard notions in non-monotonic reasoning over logic programs.
De?nition1(cf.[22,23,24]).A model M of a disjunctive program P is called well-supported if there exists a function l:B P→N such that for each A∈M there exists a rule A∨H1∨···∨H m←A1,...,A k in ground(P)with M|=A i and l(A)>l(A i)for all i and k.
De?nition1is a straightforward adaptation of the notion of well-supported model for non-disjunctive programs,as given in[23].For non-disjunctive(and negation-free)programs,the well-supported models are exactly the minimal ones,but this is not in general the case for disjunctive programs:Just consider the program consisting of the single rule p∨q←.Then{p,q}is a well-supported model,but is not minimal.
Lifted appropriately to(non-disjunctive)programs with negation,the well-supported models coincide with the well-known stable models.This was shown in [23]and studied in-depth in[24,25].Stable models[26]provide the base for the most popular non-monotonic reasoning paradigm called Answer Set Program-ming,of which the two most prominent implementations are Dlv and Smodels [27,28].Our results thus stand within this well-established tradition.
It is apparent that A∈B P is entailed by a(disjunctive)program P by approximate SLD-resolution if and only if it is true in at least one well-supported model of P.This is called brave reasoning with well-supported models.A formal proof of the following proposition is omitted for space restrictions. Proposition1.Entailment of ground conjunctive queries under approximate SLD-resolution is brave reasoning with well-supported models.
As an example,consider the(propositional)program consisting of the two rules p∨q←and r←p∧q.Its minimal models are{q}and{p},so r is not bravely entailed by reasoning with minimal models.However all of{q},{p},{p,q}and {p,q,r}are well-supported models,so r is bravely entailed by reasoning with well-supported models.
Resolution-Based Approximate Reasoning for OWL DL391 There is an alternative way of formalizing approximate SLD-resolution using a modi?ed notion of split program[29].Given a rule
H1∨···∨H m←A1,...,A k,
the derived split rules are de?ned as:
H1←A1,...,A k...H m←A1,...,A k.
For a given disjunctive program P its split program P is de?ned as the collection of all split rules derived from rules in P.Approximate SLD-resolution on P is obviously identical to SLD-resolution over P .
Minimal models are well-supported,as can be seen from the following result which was obtained along the lines of research laid out in[24,25].
Theorem2([30]).Let P be a disjunctive program.Then a model M of P is a minimal model of P if and only if there exists a function l:B P→N such that for each A which is true in M there exists a rule A∨H1∨···∨H m←A1,...,A k in ground(P)with M|=A i,M|=H k and l(A)>l(A i)for all i and k.
We hence have the following result,noting that P|=Q for any ground con-junctive query Q and program P if and only if Q is true in all minimal models of P.
Proposition2.Let P be a(possibly disjunctive)program and Q be a ground conjunctive query with P|=Q.Then there exists an approximate SLD-refutation for Q.
We remark that for negation-free disjunctive programs minimal models again coincide with answer sets[26],as in the currently evolving Answer Set Program-ming Systems,as already mentioned.
5.2Approximate Resolution for OWL DL
Our proposal is based on the idea of converting a given OWL DL knowledge base into a function-free de?nite disjunctive logic program,and then to apply approximate resolution for ABox reasoning.
In order to be able to deal with all of OWL DL,we need to add a pre-processing step to get rid of nominals,i.e.we need to compile SHOIN(D) ontologies to SHIQ(D).We can do this by Language Weakening as follows:For every occurrence of{o1,...,o n},where n∈N and the o i are abstract or con-crete individuals,replace{o1,...,o n}by some new concept name D,and add ABox assertions D(o1),...,D(o n)to the knowledge base.Note that the trans-formation just given does in general not yield a logically equivalent knowledge base,so some information is lost in the process.Putting all the pieces together, we propose the following subsequent steps for approximate ABox reasoning for OWL DL.
392P.Hitzler and D.Vrandeˇc i′c
1.Apply Language Weakening as just mentioned in order to obtain a SHIQ(D)
knowledge base.
2.Apply transformations as in Section4in order to obtain a negation-free
disjunctive datalog program.
3.Apply approximate SLD-resolution for query-answering.
The?rst two steps can be considered to be preprocessing steps for setting up the intensional part of the database.ABox reasoning is then done in the last step. From our discussions,we can conclude the following properties of approximate ABox reasoning for SHIQ(D).
–It is complete with respect to?rst-order predicate logic semantics.
–It is sound and complete wrt.brave reasoning with well-supported models.–Data complexity of our approach is polynomial.
6Screech OWL
A preliminary implementation of our approach is available as the Screech OWL approximate reasoner.2It is part of the KAON2OWL tools.3
KAON24is the KArlsruhe ONtology framework,which includes a fast OWL reasoner based on the transformation algorithms mentioned in Section4,and also includes many other features helpful to work with ontologies.Among the KAON2 OWL tools,deo performs the language weakening step described in Section5.2 in order to obtain a SHIQ(D)knowledge base.As KAON2implements the sophisticated translation algorithms described in Section4,we can convert an OWL ontology into a disjunctive datalog program,e.g.by using the dlpconvert KAON2OWL tool with the-x switch.
Screech then accesses the results of the translation through the KAON2 API,creates the corresponding split programs and serialises them as Horn logic programs in Edinburgh Prolog syntax.The result can be fed to any Prolog interpreter—or other logic programming engine—,which in turn can be used to perform ABox reasoning and inferencing over the knowledge base.
For completeness,we need to mention that in general support for concrete do-mains and other features like integrity constraints is not necessarily implemented in o?-the-shelf logic programming systems.In these cases,concrete domains etc. cannot be used.The KAON2OWL tool ded,3for example,performs a language weakening step by removing all concrete domains,and may come in handy in such situations.
7An Example
We demonstrate our approach by means of a simple OWL DL ontology.It con-tains only a class hierarchy and an ABox,and no roles,but this will su?ce to display the main issues.
2http://logic.aifb.uni-karlsruhe.de/screech
3http://www.aifb.uni-karlsruhe.de/WBS/dvr/owltools
4https://www.doczj.com/doc/2a4117938.html,
Resolution-Based Approximate Reasoning for OWL DL393
serbian croatian european
eucitizen european
german french beneluxian eucitizen
beneluxian≡luxembourgian dutch belgian serbian(ljiljana)serbian(nenad)german(pascal)french(julien)
croatian(boris)german(markus)german(stephan)croatian(denny)
indian(sudhir)belgian(saartje)german(rudi)german(york)
Fig.3.Example ontology
The ontology is shown in Figure3,and its intended meaning is self-explanatory. Note that the fourth line,
beneluxian≡luxembourgian dutch belgian,
translates into the four clauses
luxembourgian(x)∨dutch(x)∨belgian(x)←beneluxian(x),(1)
beneluxian(x)←luxembourgian(x),
beneluxian(x)←dutch,
and beneluxian(x)←belgian(x).
Thus,our approach changes the ontology by treating the disjunctions in line(1)as conjunctions.This change a?ects the soundness of the reasoning procedure.However,most of the ABox consequences which can be derived by approximate SLD-resolution are still correct.Indeed,there are only two derivable facts which do not follow from the knowledge base by classical reasoning,namely
dutch(saartje)and luxemburgian(saartje).
All other derivable facts are correct.
Screech translates the ontology from Figure3into the Prolog program listed in Figure4.As standard implementations of SLD-resolution do not use fair selection functions and also use depth-?rst search for higher e?ciency,they may sometimes fail to produce answers because they run into in?nite branches of the search tree.This occurs,for example,when using SWI-Prolog5.A reordering of the clauses may improve the results,but does not solve the problem entirely. More satisfactory performance can be obtained by using SLD-resolution with tabling,as implemented e.g.in the XSB Prolog system6.In this case,all desired consequences can be derived.
5https://www.doczj.com/doc/2a4117938.html,/
6https://www.doczj.com/doc/2a4117938.html,
394P.Hitzler and D.Vrandeˇc i′c
serbian(ljiljana).serbian(nenad).german(pascal).french(julien). croatian(boris).german(markus).german(stephan).croatian(denny). indian(sudhir).belgian(saartje).german(rudi).german(york). european(X):-serbian(X).
european(X):-croatian(X).
european(X):-eucitizen(X).
eucitizen(X):-german(X).
eucitizen(X):-french(X).
eucitizen(X):-beneluxian(X).
beneluxian(X):-luxembourgian(X).
beneluxian(X):-dutch(X).
beneluxian(X):-belgian(X).
dutch(X):-beneluxian(X).
luxembourgian(X):-beneluxian(X).
belgian(X):-beneluxian(X).
Fig.4.Example Screech output
8Experiments and Evaluation
An approximate reasoning procedure needs to be evaluated on real data from practical applications.Handcrafted examples are of only limited use as the ap-plicability of approximate methods depends on the structure inherent in the experimental data.
For our evaluation we have performed experiments with the OWL DL version of the GALEN Upper Ontology,7as it appears to be su?ciently natural and real-istic.As it is a TBox ontology only,we populated GALEN’s175classes randomly with500individuals.8GALEN does not contain nominals or concrete domains. GALEN has673axioms(the population added another500).The TBox trans-lation to disjunctive datalog took about2300ms,after which we obtained2687 disjunctive datalog rules containing267disjunctions within133rules.Among these were152integrity constraints(i.e.rules with empty head),which we re-moved for our experiment as they led to inconsistency of the database.9After splitting disjunctive rules,we arrived at2802Horn rules.
We then randomly selected classes and queried for their extension using the KAON2datalog engine,both for processing the disjunctive datalog program and for the split program.Some of the typical results are listed in Table2,which indicates a signi?cant speed-up of more than40%on average,while the vast ma-jority of the retrieved answers is correct.Note that we obtain signi?cant speed-up although the KAON2datalog engine is not optimized for Horn programs,but rather tuned to e?cient performance on de?nite disjunctive datalog.
7https://www.doczj.com/doc/2a4117938.html,/～rector/ontologies/simple-top-bio/
8Using the pop KAON2OWL tool.
9This is an expected e?ect.Removal of the integrity constraints does not destroy completeness of the approximate reasoning procedure.
Resolution-Based Approximate Reasoning for OWL DL395 Table2.Performance comparison for instance retrieval using disjunctive datalog(DD) vs.the corresponding split program(SPLIT),on the KAON2datalog engine.Instances indicates the number of instances retrieved using DD versus SPLIT,e.g.class Multi-ple contained9individuals,while the split program allowed to retrieve13(i.e.the9 correct individuals plus4incorrect ones).The full name of the class in the last row is Biological object that has left right symmetry.
Time(DD)Time(SPLIT)Instances Class Name
11036ms6489ms154/154Biological object
11026ms5959ms9/9Speci?ed set
11006ms6219ms9/13Multiple
11015ms5898ms16/16Probe structural part of heart
11036ms7711ms4/4Human red blood cell mature
11055ms5949ms24/58Biological object that...
The times were obtained with initial Java VM memory set to256MByte.Un-der memory restrictions,the speed-up is more signi?cant,which is probably caused by the necessity to allocate additional memory for the DD reasoning task.Corre-sponding?gures are given in Table3.Our experiments also indicate that Screech may be useful when hardware is limited,for example in portable devices.
9Conclusions and Further Work
In a nutshell,our proposed procedure approximates reasoning by disregarding non-Horn features of OWL DL ontologies.We argue that this is a reasonable approach to approximate reasoning with OWL DL in particular because many of the cur-rently existing ontologies rarely use language constructs that do not fall into the Horn fragment of OWL DL[31].So it can be projected that even in the future these constructs will play a minor role and thus should be the?rst to be tempered with in order to gain tractable reasoning.
Our approach provides ABox reasoning with polynomial time complexity. While it is complete,it is also unsound with respect to?rst-order logic.We have shown,however,that the inference underlying our approach can be characterized using standard methods from the area of non-monotonic reasoning.We have also presented our implementation Screech,and veri?ed the usefulness of our approach by means of experiments.
Table3.Performance comparison as in Table2,but with128MByte intial memory
Time(DD)Time(SPLIT)Instances Class Name
32997ms4817ms154/154Biological object
33028ms4947ms9/9Speci?ed set
32927ms4987ms9/13Multiple
32977ms4957ms16/16Probe structural part of heart
32987ms7350ms4/4Human red blood cell mature
32947ms4796ms24/58Biological object that...
396P.Hitzler and D.Vrandeˇc i′c
The checking whether a conjunctive query is a predicate logic consequence of a (negation-free)disjunctive logic program P amounts to checking whether the query is valid in all minimal models of P,i.e.corresponds to cautious reasoning with minimal models.Theorem2suggests how an anytime algorithm for this might be obtained:After performing approximate SLD-resolution,it remains to be checked whether there is any(ground instance of a)rule used in the refutation of the query, which has an atom A in its head besides the one used in the refutation and such that A is(cautiously)entailed by the program.Such an algorithm might then?rst?nd a brave proof of a query,and then substantiate this proof by subsequent calculations. Our approach may also be useful for the quick derivation of possible answers to a query,which may then be used for e?cient guidance of the search within a sound and complete OWL reasoner.These and other issues are currently under investigation. Acknowledgement.We are grateful for discussions with Boris Motik about his and our work.
References
1.Hustadt,U.,Motik,B.,Sattler,U.:Reasoning for Description Logics around SHIQ
in a Resolution Framework.Technical Report3-8-04/04,FZI,Karlsruhe,Germany (2004)http://www.fzi.de/wim/publikationen.php?id=1172.
2.Hustadt,U.,Motik,B.,Sattler,U.:Reducing SHIQ?Description Logic to Disjunc-
tive Datalog Programs.In:Proc.of the9th Conference on Knowledge Representation and Reasoning(KR2004),AAAI Press(2004)
3.Hustadt,U.,Motik,B.,Sattler,U.:Reasoning in description logics with a concrete
domain in the framework of resolution.In de M′a ntaras,R.L.,Saitta,L.,eds.:Pro-ceedings of the16th Eureopean Conference on Arti?cial Intelligence,ECAI’2004,in-cluding Prestigious Applicants of Intelligent Systems,PAIS2004,Valencia,Spain, August22-27,2004,IOS Press(2004)353–357
4.Motik,B.,Sattler,U.,Studer,R.:Query answering for OWL-DL with rules.In:Pro-
ceedings of the3rd International Semantic Web Conference(ISWC2004),Hiroshima, Japan,November2004.(2004)To appear.
5.W3C:Web ontology language(OWL)https://www.doczj.com/doc/2a4117938.html,/2004/OWL/(2004)
6.Baader,F.,Calvanese,D.,McGuinness,D.,Nardi,D.,Patel-Schneider,P.,eds.:The
Description Logic Handbook.Cambridge University Press(2003)
7.Hustadt,U.,Motik,B.,Sattler,U.:Data complexity of reasoning in very expressive
description logics.In Kaelbling,L.P.,Sa?otti,A.,eds.:Proceedings of the Nineteenth International Joint Conference on Arti?cial Intelligence,Edinburgh,Scotland.(2005) 466–471
8.Makinson,D.:Bridges from Classical to Nonmonotonic Logic.Volume5of Texts in
Computing.King’s College Publications,London(2005)
9.Horrocks,I.,Patel-Schneider,P.F.:A Proposal for an OWL Rules Language.In:
Proc.of the Thirteenth Int’l World Wide Web Conf.(WWW2004),ACM(2004) 10.Lutz,C.:Description Logics with Concrete Domains—A Survey.In:Advances in
Modal Logics.Volume4.,King’s College Publications(2003)
11.Horrocks,I.,Sattler,U.,Tobies,S.:Practical Reasoning for Very Expressive Descrip-
tion Logics.Logic Journal of the IGPL8(2000)239–263
12.Lloyd,J.W.:Foundations of Logic Programming.Springer,Berlin(1988)
Resolution-Based Approximate Reasoning for OWL DL397 13.Selman,B.,Kautz,H.A.:Knowledge compilation using Horn approximations.In:
Proceedings of the Ninth National Conference on Arti?cial Intelligence(AAAI-91).
(1991)904–909
14.Schaerf,M.,Cadoli,M.:Tractable reasoning via approximation.Arti?cial Intelligence
74(1995)249–310
15.Dalal,M.:Anytime clausal reasoning.Annals of Mathematics and Arti?cial Intelli-
gence22(1998)297–318
16.Cadoli,M.,Scarcello,F.:Semantical and computational aspects of Horn approxima-
tions.Arti?cial Intelligence119(2000)
17.van Harmelen,F.,ten Teije,A.:Describing problem solving methods using anytime
performance pro?les.In:Proceedings of ECAI’00,Berlin(2000)181–186
18.Groot,P.,ten Teije,A.,van Harmelen,F.:Towards a structured analysis of approx-
imate problem solving:a case study in classi?cation.In:Proceedings of the Ninth International Conference on Principles of Knowledge Representation and Reasoning (KR’04),Whistler,Colorado(2004)
19.Stuckenschmidt,H.,van Harmelen,F.:Approximating terminological queries.In
Larsen,H.,et al,eds.:Proc.of the4th International Conference on Flexible Query Answering Systems(FQAS)’02).Advances in Soft Computing,Springer(2002) 20.Horrocks,I.,Li,L.,Turi,D.,Bechhofer,S.:The Instance Store:DL reasoning with
large numbers of individuals.In:Proceedings of the International Workshop on De-scription Logics,DL2004,Whistler,Canada.(2004)31–40
21.Groot,P.,Stuckenschmidt,H.,Wache,H.:Approximating description logic classi?ca-
tion for semantic web reasoning.In G′o mez-P′e rez,A.,Euzenat,J.,eds.:The Seman-tic Web:Research and Applications,Second European Semantic Web Conference, ESWC2005,Heraklion,Crete,Greece,May29-June1,2005,Proceedings.Volume 3532of Lecture Notes in Computer Science.,Springer(2005)318–332
22.Apt,K.R.,Blair,H.A.,Walker,A.:Towards a theory of declarative knowledge.In
Minker,J.,ed.:Foundations of Deductive Databases and Logic Programming.Mor-gan Kaufmann,Los Altos,CA(1988)89–148
23.Fages,F.:Consistency of Clark’s completion and existence of stable models.Journal
of Methods of Logic in Computer Science1(1994)51–60
24.Hitzler,P.,Wendt,M.:A uniform approach to logic programming semantics.Theory
and Practice of Logic Programming5(2005)123–159
25.Hitzler,P.:Towards a systematic account of di?erent semantics for logic programs.
Journal of Logic and Computation15(2005)391–404
26.Gelfond,M.,Lifschitz,V.:Classical negation in logic programs and disjunctive
databases.New Generation Computing9(1991)365–385
27.Eiter,T.,Leone,N.,Mateis,C.,Pfeifer,G.,Scarcello,F.:A deductive system for
nonmonotonic reasoning.In Dix,J.,et al,eds.:Proceedings of the4th International Conference on Logic Programming and Nonmonotonic Reasoning(LPNMR’97).Vol-ume1265of Lecture Notes in Arti?cial Intelligence.,Springer,Berlin(1997)
28.Simons,P.,Niemel¨a,I.,Soininen,T.:Extending and implementing the stable model
semantics.Arti?cial Intelligence138(2002)181–234
29.Sakama,C.,Inoue,K.:An alternative approach to the semantics of disjunctive logic
programs and deductive databases.Journal of Automated Reasoning13(1994) 145–172
30.Knorr,M.:Level mapping characterizations for quantitative and disjunctive logic
programs.Bachelor’s Thesis,Department of Computer Science,Technische Univer-sit¨a t Dresden,Germany(2003)
31.Volz,R.:Web Ontology Reasoning with Logic Databases.PhD thesis,AIFB,Uni-
versity of Karlsruhe(2004)
佳能500D功能操作详解 虽说是入门级的单反相机，但瘦死骆驼比马大，在画面的成像质量和快门速度上还是要远远的强于卡片机。不过话又说回来，即使是入门级的卡片机对于本人这种初级的新手来说还是一个看似不可琢磨的高深玩意，想必不少朋友也有笔者这种情况，对于入门级的单反相机的使用还是存在些问题。而这些问题虽不是很深奥，但对于使用过程中的影响还是很大的，针对这样的情况，笔者就将在初使用单反的过程中遇到的一些问题和解决办法来给大家简单的介绍下。 对于刚刚入门的新手来说来用得比较多的也就是快门、光圈、ISO感光度这三个要素。要很好的应用这三个要素首先就得了解其具体作用是什么，下面笔者抱着和大家一块学习的心态来就这三个要素给大家详细的说说。 一：快门，配合光圈的变化，可以调整快门的速度来实现正确曝光，快门就是曝光时间的长短，比如你的光圈确定为F8，那么快门越快，进来的光亮就越少，快门越慢就进光更多，快速的快门可以把运动瞬间凝结在底片或者CCD上。可以举个形象的比喻，相机的镜头就像人的眼睛，快门就是人的眼皮，快门越快，就是眼睛眨得越快，而眼皮眨得越快相对在两个眨眼之间那个瞬间接收的光线越少，就会觉得很暗；如果眨的越慢，在两个眨眼之间的接收光线越多，而显得光线充足。 不过快门的快慢还会影响到一个方面，就是快门快的话相对就会减少画面糊的情况发生，而快门很慢可能会导致画面很糊，这个道理也很简单，想想卡片机为什么会有那么多的糊的照片就明白了，就是快门速度的原因。所以要寻求一个既可以采集到充足的光线而又不会导致轻微的手动就会影响到画面质量的平衡点。 二：光圈，镜头的通光量有光圈的控制作用，光圈值越大，（既F值越小），通过的光亮就越多，大光圈带来的特点就是能够获得很浅的景深，大光圈的效果对于拍摄人物很适合，能够突出主体。当然，大光圈下面的聚焦一定要保持准确，否则比较浅的景深很容易照成焦点的偏差。光圈越小（既F值越大），通过的光亮就越少，在小光圈下面可以获得比较长的景深，小光圈的设计很适合拍摄风景或是建筑一类的物体。 三：ISO值，它是胶片或CCD感光能力的大小。正常卡片机在白天室内外一般都是用ISO800或是ISO1600这两个值，如果要画面更细腻的话，可以通过降低ISO值来获得，这样的设置适合于拍摄人像或风光静物等场景，层次非常丰富。然而在单反相机中，低感光度带来的影像是造成感光时间加长，不得不使用放大光圈或者放慢快门来补充曝光，以达到正确的画面要求。如果感光度值很高，那么就会带来胶片画面的颗粒感就会增强，画面就有噪点产生，它的好处在于，可以让整个画面更佳的明亮，同时也可以选择更快的快门速度或者更小的光圈。这种方式比较适合抓拍运动场面或者动态景物，合理运用可以产生比较特殊的效果。在比较暗的环境下面，提高感光度值也是一个好办法，不过对于要求画质的效果的朋友来说还是尽量少用提高ISO值的方法。 而在平时的使用过程中笔者发现一般光圈和快门的调节比较多，而且很讲究二者的搭配使用，那么究竟如何搭配这两个应用的值呢。
佳能摄像机说明书 篇一：佳能相机说明书 佳能相机说明书使用 EOS 500D 拍摄短片的简易步骤最近更新 : 15-May-20XX 发行编号 : 820XX98200解决方法本页介绍如何以默认设置拍摄短片。关于如何更改设置拍摄短片，请参阅相关信息。须知拍摄短片时，请使用 SD 速率级别 6 “储卡，短片拍摄和回放可能不正常工作。此外，建议使用大容量的存储卡进行短片拍摄。 1. 将相机电源开关置于。 2. 将镜头对焦模式开关设为。”或更高速的存储卡。如果使用 SD 速率级别较低的存3. 将模式转盘设为。反光镜会发出声音，然后图像会出现在液晶监视器上。 4. 将显示自动对焦点。可以按下将自动对焦点移动到您想要对焦的位置。? 无法移动到图像的边缘。 ? 若要让自动对焦点返回中央，按下按钮。 5. 将自动对焦点对准主体，然后按住按钮。 6. 按下按钮开始拍摄短片。”标记将显示在屏幕的右上方。 7. 再次按下停止拍摄短片。注意短片拍摄期间，请不要把镜头对准太阳。太阳的热量会损坏相机内部部件。注释记录和图像画质? 短片文件格式将为 MOV。 ? 如果单个短片的文件尺寸达到 4GB，短片拍摄将自动停止。若要重新开始短片拍摄，按下按钮。（开始录制新的短片文件。）? 相机的内置麦克风还会录制相机操作音。如果不想录制任何
声音，将声音录制设置为［关］。回放和电视机连接? 如果用 HDMI 电缆将相机连接到电视机并以［1920x1080］或［1280x720］拍摄短片，将以较小的尺寸在电视机上显示所拍摄的短片。然而，实际的短片将以所设置的短片记录尺寸正确录制。? 如果将相机连接到电视机并拍摄短片，在拍摄期间电视机将不会输出任何声音。然而，声音将被正确录制。须知 ? 如果长时间连续拍摄短片，相机的内部温度会升高。这会降低图像画质。不拍摄时，请关闭电源。? 如果在直射阳光或高温条件下进行拍摄，屏幕上可能会出现图标（内部高温警告）。如果在显示警告图标的状态下继续拍摄短片，图像画质可能会降低。应该关闭电源并让相机休息片刻。? 如果在显示图标的状态下继续拍摄短片直到相机的内部温度进一步升高，短片拍摄将会自动停止。如果发生这种情况，相机的内部温度降低前，您将无法再次进行拍摄。关闭电源并让相机休息片刻。 设置白平衡最近更新 : 15-May-20XX 发行编号 : 820XX95000解决方法购买时，佳能数码相机和便携式数码摄像机的白平衡设置为“自动白平衡”。设置通常将获取正确的白平衡。如果用不能获得自然的色彩，可以选择用于不同光源的白平衡设置或通过拍摄白色物体手动设置白平衡。注释在基本拍摄区模式中，将自动设置为。要设置白平衡，请遵循以下步骤。 1. 将电源开关设为。 2. 相机
合格骨健康基本补充剂—绝代双骄之钙剂与维生素D 100分 1.有机酸钙实际吸收率是（A) A.4-30% B.10-30% C.20-30% D.20-50% 2.建议定期监测患者血清钙及12h尿钙浓度，根据其水平调整药物剂量，以避免药物过量所引发的高钙血症或高尿钙的发生。（B) A.正确 B.错误 3.咖啡因摄入过多可增加尿钙的排出，还能轻度减弱（)钙吸收，同时对钙储留有中度影响，并与增加女性镜骨骨折危险性有联系。（C) A.胃 B.骨骼 C.肠
4.补充维生素D时，应首选维生素D3,如果不能得到维生素D3时，可用维生素D2代替。（A) A.正确 B.错误 5.长期使用活性VD治疗的患者，建议在启动治疗后的1、3、6个月分别监测尿钙磷及血钙磷水平，此后建议每年监测两次血钙磷、尿钙磷及肾功能。（A) A.正确 B.错误 6.VD缺乏的矫正非紧迫时，或当补充VD同时口服抗骨吸收药物，或许可以进行维持治疗方案，而无需使用负荷剂量。（A) 。A.正确 B.错误 7.下述给药方法已证明无效，或无效和中毒的风险增加，因此推荐，①采用肌肉注射或口服维生素D的年储备给药方法；②使用活性维生素D制剂（骨化三醇和阿法骨化醇）纠正VD缺乏。（A)
B.错误 多选题：每道题有两个或两个以上的答案，多选漏选均不得分。 1.下列属于骨质疏松的是（ABCD) A.充足的钙和维生素D的摄入对于防治骨质疏松症至关重要 B.充足的钙摄入对获得理想骨峰值，减缓骨丢失，改善骨矿化和维护骨骼健康有益。 C.充足的维生素D可增加肠钙吸收、促进骨骼矿化、保持肌力、改善平衡能力和降低跌倒风险 D.尽可能通过饮食摄入充足的钙，饮食中钙摄入不足时，可给予钙剂补充 2.如何防晒？(ABC) A含氧化锌和二氧化钛等成分的“广诺防晒霜 B.干爽的胜过潮湿的布料 C.限制接触紫外线，太阳最强时戴太阳镜和帽子 D.面部、双上肢和双下肢暴露于阳光5-30min
补钙为什么要补充维生素d 钙和维生素d是我们人体不可分割的，在当今时代都是这样的。如今的父母都是希望自己的孩子能够茁壮的成长，所以一定要在补钙的时候也是要增加维生素d这样才能更有效的促进人 体的吸收。那么，补钙为什么要补充维生素d？ 1、先了解下钙和维生素D对人体的作用 钙是构成人体的重要组成部分，占人体重量的百分之1.5到百分之2，其中百分之99存在骨骼和牙齿中。钙对维生素神经兴奋性、血液凝固、肌肉收缩和舒张、腺体的分泌有着很总要的作用，并对多种酶有激活作用。钙主要受甲状旁腺素、降钙素与1.25-（OH）D3的共同调节，钙的吸收也需要一定的磷，磷缺乏也会影响钙的吸收。 维生素D主要与甲状旁腺共同作用，维持血钙的水平稳定，它是钙磷代谢重要的调节因子之一，维持着钙和磷的正常水平，
这对正常骨骼的矿化、肌肉收缩、神经传导以及体内所有细胞的功能都是必须的，同时维生素D还具有免疫调节功能，可以改变机体对感染的反应。 2、婴儿时期需要补充钙剂和维生素D吗？ 答案是肯定的，婴儿时期是处于快速生长发育期，对维生素D的需求来那个相对较大，母乳中的维生素D的水平比较低，维生素D既可以由膳食供给，又可以经暴露日光的皮肤合成。如果小婴儿缺乏钙同时也缺乏维生素D，那么就会出现骨质软化症和佝偻病，表现出多汗、易惊生长手足抽搐。久而就会出现骨质软化便会导致乒乓球头颅，肋缘外翻、串珠、手镯、X或者O型腿。 3、每个时期的婴幼儿该如何添加钙剂和维生素D呢？ 一般来说，0-6个月的孩子，纯母乳喂养每天需要钙三百毫克，人工喂养的婴儿考虑到非母乳来源钙吸收率比较低，固定为每天四百毫克。这时候的维生素D的含量为固定的一天一粒，也
数码相机用机宝典是PConline数码相机频道为解决网友用机疑惑的新栏目，帮助初涉摄影的网友逐步掌握摄影技巧，拍摄出好照片。用机宝典每次针对一款机型，从基本的参数设置到进阶使用，还会包括各种相关附件的使用，逐一讲述拍摄技巧和需要注意的要点，力求通过深入浅出的讲解帮助网友进步。本期用机宝典则是围绕相当热门的入门单反佳能500D进行拍摄技巧的讲解，欢迎大家对用机宝典提出自己的见解，帮助我们做得更好。 步骤/方法 1.一、佳能500D简介 2. 佳能500D是佳能门内最新的入门级数码单反，习惯小改小闹 推出升级产品的佳能，这次在500D身上的做法反而十分豪爽，让不少其他门派的粉丝也羡慕不已。佳能500D可谓是一个50D
和5D Mark II（无敌兔）的混合版，它采用了50D身上的1510万像素CMOS、DIGIC 4图像处理引擎、3.0英寸92万像素LCD 和其余的一些特色功能，更拥有在无敌兔身上的1080p高清摄像能力，这样一个混合产品注定出现以下犯上的局面。 3. 不过，作为佳能入门级的机型，在外观设计上仍然是廉价的做法，另外在对焦系统和快么寿命上也有所不同，仍然符合其入门级的定位。这么一次大跃进的升级，使得佳能500D具有了空前丰富的功能，下面我们一起来看看： 4.二、佳能500D基础使用指南 5.1、怎样使用CA模式
6. CA模式也就是创意自动功能，它推出的目的在于照顾入门级的使用者，使得普通用户也可调整图像亮度，或进行背景虚化等。 使用它的时候，只需要把模式转盘转到“CA”模式即可。这时如果不进行任何调节的话，相机基本上会按照P档的情况进行处理。 7. 不过，相机的判断不一定会按照人的意志的，这时则可以通过
佳能500D使用技巧整理 个人整理 使用佳能500D拍摄各种照片 全部内容如下： 1 如何正确暴光 2 光圈快门速度ISO之间的关系 3 如何拍背景虚化的照片 4 镜头焦距的选择 5 如何拍夜景 6 被摄体不在画面中间时如何对焦 7 如何在黑天室外拍人像 8 如何拍摄动态汽车照片 9 如何拍摄鬼影照片 10 如何拍摄剪影照片 11 如何拍摄烟花 12 婚礼上如何拍照 13 台球室拍照 14 如何拍显个，如何拍显瘦，哪个角度最漂亮。。 注意：和我学可学不成大师，很可能还学些凑毛病以后不好改，不过按我说的做能在短期内在外行面前冒充冒充高手。。 现在开始按我的套路来了。。。。。。 1 如何正确暴光 先把相机拍摄模式调到M档，也就是全手动档。。（如果不明白如何调到M档就看一下说明书，以后就不再说明了，看不懂的地方就翻说明书），有人说M档不实用操作麻烦，新手不容易掌握，但我觉得M档更适合新手学习，因为M档需要手动调圈和快门更容易让新手快速的掌握光圈和快门的关系以及对成像的影响。。反正我拍照只要不是很紧急的情况下拍摄都是M档的。。 要想用M档拍照先要知道如何调光圈和快门，从液晶屏上看横排最上方有三个数据，最左边的是快门速度，中间是光劝大小，右边是ISO感光度值。。正常在光线不是很暗的情况下我们拍照都要把ISO值设成一百，如果光线很暗又不想开闪光灯的话可以适当调高。。然后光圈和速度怎么调这个是不固定的，要看你所在的环境光线和你要拍摄的物体。。别急，一步步来…… 说一下M档下如何调快门速度光圈和ISO，直接转动控制转盘就可以调整快门速度，按住液晶屏右上角那Av按钮的同时转动转盘就是调光圈大小，控制转盘边上有个ISO按钮，按一下然后用机背的方向键选择ISO，选好后按方向键中间的SET按钮确定。。 知道怎么调以后我们就要实战演练了，先找个布娃娃或别的差不多的东西或是家里的任意一样摆设，先拿它练起。。这时候你要拿媳妇练估计容易给你媳妇练出拍照恐惧症~正常情况下18-55IS镜头最佳成像光圈是8。。先把光圈设成8再说。。这时候的设置是光圈8，ISO100。。设置好后把眼睛贴近取景窗，然后瞄准被摄体，半按快门，听到嘀一声响表示对焦完成就可以拍摄了， 在取景器里我们能看到下方有这样一排发绿光的数字和标尺，先说一下这分别是什么意思，最左边的是
考试成绩100分，考试通过 骨健康基本补充剂—绝代双骄之钙剂与维生素D 单选题：每道题只有一个答案。 1.长期使用活性VD治疗的患者，建议在启动治疗后的1、3、6个月分别监测尿钙磷及血钙磷水平，此后建议每年监测两次血钙磷、尿钙磷及肾功能。（） A.正确 B.错误 2.建议定期监测患者血清钙及12h尿钙浓度，根据其水平调整药物剂量，以避免药物过量所引发的高钙血症或高尿钙的发生。（） A.正确 B.错误 3.下述给药方法已证明无效，或无效和中毒的风险增加，因此推荐，①采用肌肉注射或口服维生素D 的年储备给药方法；②使用活性维生素D 制剂(骨化三醇和阿法骨化醇)纠正VD 缺乏。（） A.正确 B.错误 4.咖啡因摄入过多可增加尿钙的排出，还能轻度减弱（）钙吸收，同时对钙储留有中度影响，并与增加女性髋骨骨折危险性有联系。 A.胃 B.骨骼 C.肠 D.血液 5.有机酸钙实际吸收率是（）
A.4-30% B.10-30% C.20-30% D.20-50% 6.补充维生素D时，应首选维生素D3，如果不能得到维生素D3时，可用维生素D2代替。（） A.正确 B.错误 7.VD缺乏的矫正非紧迫时，或当补充VD同时口服抗骨吸收药物，或许可以进行维持治疗方案，而无需使用负荷剂量。（） A.正确 B.错误 多选题：每道题有两个或两个以上的答案，多选漏选均不得分。 1.如何防晒？（） A.含氧化锌和二氧化钛等成分的“广谱”防晒霜 B.干爽的胜过潮湿的布料 C.限制 接触紫外线，太阳最强时戴太阳镜和帽子 D.面部、双上肢和双下肢暴露于阳光5-30 min 2.普通维生素D是否需要常规监测（） A.普通VD安全剂量范围宽 B.没必要常规监测血清25OHD C.通过血清25OHD 判断是否中毒 D.有维生素D缺乏症状、吸收不良、怀疑用药效果不佳等，可能需要监 测血清25OHD 3.下列属于骨质疏松的是（）
用佳能5d2单反摄像技巧及注意事项——很详细5d2在电影领域中的应用电影《终结者2018》的摄影指导 shanehurlbertasc美国电影摄影家协会会员他用5d2拍的短片《最后三分钟》在摄影界引起了轩然大波单反视频早已成为广大发烧友和专业影视从业人员热捧的新鲜事物了。5d2的变革就在于用低廉的价格能让普通大众拥有最好的甚至超越35mm电影摄影机的动态画质a:硬件和配套设备部分1三角架极其重要。照相机的架子本身不是为移动拍摄的所以最好找摄像机脚架这样才可以平稳拍摄摇画面也才能保证长焦拍摄的画面稳定。例如曼富图利拍libec的摄像机专业架。唯一小问题是摄像机托板装在照相机上使用时间长后照相机下面会有难看的挤压痕迹。2镜头选择。一般来说最常用的是超广角和长焦中焦段一定要大光圈定焦。因为5d2是全画幅所以可以按照电影头来选择。建议配置如下10mm或12mm--24mm超广角24--105mm光圈大一点好比如 2.824-70的佳能红圈120-500600mm炮筒50mmf1.2这个强烈推荐人像专用。3聚焦问题。目前来说如果没有合适的跟焦器配套用5d2来拍摄动态需要跟焦的画面是相当麻烦。照相机镜头的对焦阻尼是很涩的手感不好。当然了偶尔用用问题不大。但是小液晶屏幕看对焦只能看个大概。5d2有hdmi接口可以带一个720p的带hdmi的监视器这样现场显示就会比较靠谱。当然如果是专业跟焦员他看刻度的那种也行。 不过佳能镜头的刻度都在镜头的。上面非常的不方便。4电池问题。5d2的原装电池一般可以断续拍摄1张8g卡大概12分钟建议用大容量的储存卡。广告上说的可以连续拍摄2小时是直接打开机器拍的理论最大值大部分时间你还需要取景、找焦点、找画面感觉所以不可能用到这么长时间。所以电池极其重要至少要准备3快电池才能保证不会出问题。5如果经常拍摄城市宣传片或者风光片的朋友肯定会常用到定格拍摄比如日出日落夜景。这些镜头用5d2拍简直爽到极点了。一来分辨率超爽后期可以调整的空间巨大。另一点快门时间可以放的很慢夜景尤其是车流的效果非常出色。慢门能大幅度降低画面躁点画面显得特别的干净漂亮。这里特别要推荐买一个定时快门手柄可以设置间隔时间自动拍摄。6准备足够的滤色片。nduv必不可少保护镜头必须灰片也要带足因为正午阳光下iso100光圈要收到8-11而视频又不能过多加快快门速度会造成缺少运动模糊导致播放有抖动感你想要得浅景深效果就没了。nd0.30.60.91.2最好都备上。常用的可能就是0.6和0.9.7取景器超低和超高机位的大广角如果没有外置取景器那是要命的。5d2套件概念简介5d2套件通常是指用于佳能5dmark2单反相机的摄像承托系统套组设备。随着数码科技的日新月异发展canon等公司最新出品的全尺寸ccd专业单反相机 5d2.7d.50d.60d.70d.500d都拥有了高清hd1920x1080p摄 像功能.其摄影能力已经远远超越了传统磁带摄像机。拍摄视频画面质量也在逐渐接近电影摄影机。并凭借其强大的镜头兼容性和广泛的镜头源.完全可以满足宣传篇片.电视广告片.电影短片的拍摄任务。但单反照相机拍摄高清视频的缺点也暴漏出来了。虽然拍摄的画面可以和电影摄影机和摄像机有一比。相机机身设计造成了拍摄视频的不方便机身小不能肩扛通过手持又不能保持很好的画面平衡针对这一点让广大单反视频摄像爱好者对5d2等机型是又爱又恨.爱的是拍摄视频画质好更换镜头方便恨的是拍摄不方便固定镜头画面还好说可以放在三角架上拍摄.但运动起来就麻烦了总不能用手端着拍吧。就算端着拍短时间可以但长时间实在承受不了。还有就是单反相机相机的遮光罩太小在拍摄特定场景时。光线总是斜射入镜头内。让原本漂亮的画面曝光过度。还有就是专业单反拍视频时无法进行手动对焦。必定没有类似为摄像机而设计的电动伺服器。对于以上缺点遗憾不在机器本身.5d2等机器能拍摄出足够干掉广播级别摄像机的画面已经足够对得起我们了。必定5d2等机器才1万元左右.而同级别的redone.arri等专业摄影机也都在30万到400万不等.即使拍摄画面不如5d2等机器的sonyax200e190p250p松下103等高清机也价格都在2万元以上并且还都不可更欢镜头。看
佳能500d使用说明书 佳能500d使用说明书 佳能500d使用说明书 篇一: 佳能500D PAS佳能500D PAS-C数码单反详细参数，如下: 主要性能发布日期机身特性操作方式传感器类型 201X年03月 APS-C规格数码单反全手动支持 CMS 长宽比: 3:2 传感器件尺寸 2 2.3×1 4.9mm 传感器描述有效像素数影像处理器最高分辨率除尘功能: 自动，手动，添加除尘数据 1510万 DIGIC 4 4752×3168 大: 约1510万像素(4752×3168) 中: 约800万像素(3456×2304) 图像分辨率小: 约370万像素(2352×1568) RA: 约1510万像素(4752×3168) 全高清(1080P) 高清摄像镜头特点佳能EF系列镜头(包括EF-S系列镜头)(35mm换算焦距约为镜头焦距的 1.6 镜头说明倍) 类型: TTL辅助影像重合，相位检测自动对焦点: 9个自动对焦点测光范围: EV -0.5-18(23?，73?，IS 100) 对焦方式对焦模式: 单次自动对焦，人工智能伺服自动对焦，人工智能自动对焦，手动对焦(MF) 自动对焦辅助光: 由内置闪光灯发出的短促连续闪光 9点变焦环 3英寸类型:
TFT彩色液晶监视器(可分7级调节亮度) 监视器尺寸和点数: 3英寸，约92万点(VGA) 对焦点数变焦方式显示功能显示屏尺寸像素及类型视野率: 约100% 界面语言: 25种(含简体中文) 可旋转液晶屏否触摸屏是否双屏取景器不支持 否光学类型: 眼平五面镜视野率: 垂直/水平方向约95% 放大倍率: 约0.87倍(-1m-1，使用50mm镜头对无限远处对焦) 眼点: 约19mm(自目镜透镜中央起-1m-1) 取景器描述内置屈光度调节: - 3.0-+ 1.0m-1(dpt) 对焦屏: 固定式，精确磨砂反光镜: 快回型景深预视快门性能快门类型快门速度闪光灯闪光类型外接闪光灯(热靴) 电子控制纵走式焦平面快门 1/60-1/4000秒(全自动模式)，闪光同步速度1/200秒 30-1/4000，B快门(总快门速度范围。可用范围随拍摄模式各异。) 外接支持外接闪光灯: EX系列闪光灯(能用相机设置功能) 闪光模式闪光测光: E-TTL II自动闪光闪光曝光补偿: ?2级间以1/3或1/2级为单位调节闪光曝光锁闪光范围闪光指数闪光灯回电时间曝光控制 17mm镜头视角 13/43(IS 100，以米/英尺为单位) 约3秒程序自动曝光P(全自动，人像，风光，微距，运动，夜景人像，闪光灯关闭，创意自
骨健康基本补充剂绝代双骄之钙剂与维生素D 一、单选题 1、有机酸钙实际吸收率是（A）。A、4-30% 2、咖啡因摄入过多可增加尿钙的排出，还能轻度减弱（C）钙吸收，同时对钙储留有中度影响，并与增加女性髋骨危险性有联系。C、肠 3、补充维生素D时，应首选维生素D3，如果不能得到维生素D3时，可用维生素D2代替。（A）A、正确 4、下述给药方法已证明无效，或无效和中毒的风险增加，因此推荐，①采用肌肉注射或口服维生素D的年储备给药方法；②使用活性维生素D制剂（骨化三醇和阿法骨化醇）纠正VD缺乏。（B）B、错误 5、VD缺乏的矫正非紧迫时，或当补充VD同时口服抗骨吸收药物，或许可以进行维持治疗方案，而无需使用负荷剂量。（A）A、正确 6、建议定期监测患者血清钙及12h尿钙浓度，根据其水平调整药物剂量，以避免药物过量所引发的高钙血症或高尿钙的发生。（B）B、错误 7、长期使用活性VD治疗的患者，建议在启动治疗后的1、3、6个月分别监测尿钙磷及血钙磷水平，此后建议每年监测两次血钙磷、尿钙磷及肾功能。（A）A、正确 二、多选题 1、如何防嗮？（ABCD） A、含氯化锌和二氧化钛等成分的“广谱”防嗮霜 B、干爽的胜过潮湿的布料 C、限制接触紫外线，太阳最强时戴太阳镜和帽子 D、面部、双上肢和双下肢暴露于阳光5-30min 2、下列属于骨质疏松的是（ABCD）。 A、充足的钙和维生素D的摄入对于防治骨质疏松症至关重要 B、充足的钙摄入对获得理想骨峰值，减缓骨丢失，改善骨矿化和维护骨骼健康有益 C、充足的维生素D可增加肠钙吸收、促进骨骼矿化、保持肌力、改善平衡能力和降低跌倒风险 D、尽可能通过饮食摄入充足的钙，饮食中钙摄入不足时，可给予钙剂补充 3、普通维生素D是否需要常规监测（ABCD）。
执业药师继续教育合格骨健康基本补充剂—绝代双 骄之钙剂与维生素D LELE was finally revised on the morning of December 16, 2020
合格骨健康基本补充剂—绝代双骄之钙剂与维生素D 100分 1.有机酸钙实际吸收率是（A) A.4-30% B.10-30% C.20-30% D.20-50% 2.建议定期监测患者血清钙及12h尿钙浓度，根据其水平调整药物剂量，以避免药物过量所引发的高钙血症或高尿钙的发生。（B) A.正确 B.错误 3.咖啡因摄入过多可增加尿钙的排出，还能轻度减弱（)钙吸收，同时对钙储留有中度影响，并与增加女性镜骨骨折危险性有联系。（C) A.胃 B.骨骼 C.肠 D.血液 4.补充维生素D时，应首选维生素D3,如果不能得到维生素D3时，可用维生素D2代替。（A) A.正确
B.错误 5.长期使用活性VD治疗的患者，建议在启动治疗后的1、3、6个月分别监测尿钙磷及血钙磷水平，此后建议每年监测两次血钙磷、尿钙磷及肾功能。（A) A.正确 B.错误 6.VD缺乏的矫正非紧迫时，或当补充VD同时口服抗骨吸收药物，或许可以进行维持治疗方案，而无需使用负荷剂量。（A) 。A.正确 B.错误 7.下述给药方法已证明无效，或无效和中毒的风险增加，因此推荐，①采用肌肉注射或口服维生素D的年储备给药方法；②使用活性维生素D制剂（骨化三醇和阿法骨化醇）纠正VD缺乏。（A) A.正确 B.错误 多选题：每道题有两个或两个以上的答案，多选漏选均不得分。 1.下列属于骨质疏松的是（ABCD) A.充足的钙和维生素D的摄入对于防治骨质疏松症至关重要
佳能500D入门教材（浅显易懂、菜鸟必看哦） 如何使用佳能500D拍摄各种照片 前几天帮朋友买了个500D，他没接触过单反一时不知从何下手，所以给他写个简易教程，也同时请熟悉单反熟悉500D的色友给予指正，因为我拍照也是瞎拍，所以有时候说的难免有错误或不科学的地方，大家发现问题一定要及时给予究正。 买来新相机无论是老手还是新手都应当仔细阅读一遍说明书，然后再拿起相机亲自操作拍摄几天，然后回过头来再看一遍说明书，然后再拍几天，再看一遍说明书，一般至少要看三遍吧。。。反正我是第五次看说明书的时候还能从中学到新东西。因为说明书里有些容在你没有实际拍摄经验的情况下是看不懂的，所以说明书一定要多看，说明书是最好的教材。 全部容如下： 1 如何正确曝光 2 光圈快门速度ISO之间的关系 3 如何拍背景虚化的照片 4 镜头焦距的选择 5 如何拍夜景 6 被摄体不在画面中间时如何对焦
7 如何在黑天室外拍人像 8 如何拍摄动态汽车照片 9 如何拍摄鬼影照片 10 如何拍摄剪影照片 11 如何拍摄烟花 12 婚礼上如何拍照 13 台球室拍照 14 如何拍显个，如何拍显瘦，哪个角度最漂亮。 注意：和我学可学不成大师，很可能还学些凑毛病以后不好改，不过按我说的做能在短期在外行面前冒充冒充高手 现在开始按我的套路来了。。。 1 如何正确曝光 先把相机拍摄模式调到M档，也就是全手动档。（如果不明白如何调到M档就看一下说明书，以后就不再说明了，看不懂的地方就翻说明书），有人说M档不实用操作麻烦，新手不容易掌握，但我觉得M档更适合新手学习，因为M档需要手动调圈和快门更容易让新手快速的掌握光圈和快门的关系以及对成像的影响。反正我拍照只要不是很紧急的情况下拍摄都是M档的。 要想用M档拍照先要知道如何调光圈和快门，从液晶屏上看横排最上方有三个数据，最左边的是快门速度，中间是光劝大小，右边是ISO感光度值。正常在光线不是很暗的情况下我们拍照都要把
佳能400D操作说明书 数码相机用机宝典是PConline数码相机频道为解决网友用机疑惑的新栏目，帮助初涉摄影的网友逐步掌握摄影技巧，拍摄出好照片。用机宝典每次针对一款机型，从基本的参数设置到进阶使用，还会包括各种相关附件的使用，逐一讲述拍摄技巧和需要注意的要点，力求通过深入浅出的讲解帮助网友进步。第六期购机宝典则是围绕佳能元老级数码单反400D进行拍摄技巧的讲解，欢迎大家对本栏目提出自己的见解以及希望获得的讯息，帮助我们做得更好。 佳能400D简介 佳能EOS 400D(Rebel XTi)(配18-55mm镜头套机
佳能400D是佳能上一代入门数码单反相机的代表作，相比较更为低端的350D有较大的改进，有不错的实用性能，并能使用佳能全部镜头。但由于定位低端，机身采用塑料材质，握持手感仍显得不够舒适。从上市开始到退市仍不能摆脱塑料拖鞋的称号，自推出佳能450D之后，400D逐步退出市场，但发行两年的时间里，销售量巨大，至今仍有相当大的保有量。 佳能400D装备1010万像素APS-C格式CMOS，有超声波除尘功能，可明显减少自行清洁的次数。9点宽区自动对焦，使用的DIGIC II 芯片图像处理，画质上仍有不错的表现。使用2.5英寸LCD显示屏，启动速度是0.2秒，连拍速度最快3张/秒，一次最多拍摄27张JPEG 或10张RAW照片，快门速度的范围是1/4000-30秒，最高闪光同步速度为1/200秒。在市场上混了两年的佳能400D早已成为单反用户耳熟能详的一个型号。 一、佳能400D使用常识 佳能400D作为一台上市售卖两年的入门级单反，没有太多的辅助功能，所以在使用上没有太多的操作难点，相较其升级版本450D，按键排布在显示屏左边一排，各项功能也较易操实现。这里主要讲解一下刚接触单反、还不了解相关操作的买家容易忽视的问题，内容通俗易懂，相信大家一看就懂。另外，说一下数码单反维护的一些方法。 1、模式切换转盘与拨轮的使用方法
佳能500D入门教你如何正确曝光 前几天帮朋友买了个500D，他没接触过单反一时不知从何下手，所以给他写个简易教程，也同时请熟悉单反熟悉500D的色友给予指正，因为我拍照也是瞎拍，所以有时候说的难免有错误或不科学的地方，大家发现问题一定要及时给予究正。 买来新相机无论是老手还是新手都应当仔细阅读一遍说明书，然后再拿起相机亲自操作拍摄几天，然后回过头来再看一遍说明书，然后再拍几天，再看一遍说明书，一般至少要看三遍吧。。。反正我是第五次看说明书的时候还能从中学到新东西。因为说明书里有些内容在你没有实际拍摄经验的情况下是看不懂的，所以说明书一定要多看，说明书是最好的教材。 如何正确曝光 先把相机拍摄模式调到M档，也就是全手动档。（如果不明白如何调到M档就看一下说明书，以后就不再说明了，看不懂的地方就翻说明书），有人说M档不实用操作麻烦，新手不容易掌握，但我觉得M档更适合新手学习，因为M档需要手动调圈和快门更容易让新手快速的掌握光圈和快门的关系以及对成像的影响。反正我拍照只要不是很紧急的情况下拍摄都是M档的。 要想用M档拍照先要知道如何调光圈和快门，从液晶屏上看横排最上方有三个数据，最左边的是快门速度，中间是光劝大小，右边是ISO感光度值。正常在光线不是很暗的情况下我们拍照都要把ISO值设成一百，如果光线很暗又不想开闪光灯的话可以适当调高。然后光圈和速度怎么调这个是不固定的，要看你所在的环境光线和你要拍摄的物体。别急，一步步来…… 说一下M档下如何调快门速度光圈和ISO，直接转动控制转盘就可以调整快门速度，按住液晶屏右上角那Av按钮的同时转动转盘就是调光圈大小，控制转盘边上有个ISO 按钮，按一下然后用机背的方向键选择ISO，选好后按方向键中间的SET按钮确定。 知道怎么调以后我们就要实战演练了，先找个布娃娃或别的差不多的东西或是家里的任意一样摆设，先拿它练起。这时候你要拿媳妇练估计容易给你媳妇练出拍照恐惧症~正常情况下18-55IS镜头最佳成像光圈是8。先把光圈设成8再说。这时候的设置是光圈8，ISO100。设置好后把眼睛贴近取景窗，然后瞄准被摄体，半按快门，听到嘀一声响表示对焦完成就可以拍摄了， 在取景器里我们能看到下方有这样一排发绿光的数字和标尺，先说一下这分别是什么意思，最左边的是速度，画面中是1/125秒，第二个是光圈8，中间的标尺就是测试暴光度的，右边的是ISO100，我们注意观察这个暴光度标尺下面有一个小竖条，这个条的位置有时候偏左，有时候偏右，有时候在中间。偏左说明暴光不足，这时候拍出来的照片会发黑，偏右说明暴光过度，这时候拍出来的照片会发亮，只有在正中间的时候拍摄才是最合适的，但这也不是绝对的，这只是个参考，因为测光模式不同和你的对焦点都会对这有一定影响，但咱先不管它。这时候如果你的测光尺的浮标正好在中间那么你就可以直接拍摄了，如果在右边那么你就需要把速度调快，调成1/160 1/200 1/250 1/320 1/400直到标尺到正中间为止，但需要注意的是别太快了，如果你是拍静止不动的物体（包括风景静止的人物体）那就别调太快，如果到了1/500还是偏右那说明你现在正在一个阳光充足光线非常良好的环境下拍摄，这时候你就把速度停在1/500秒这然后去调整光圈把光圈调小些，把光圈调小就是指把光圈的数值调大的意思。然后一直缩小直至测光尺浮标到中间为止，如果你光圈调到22了还是偏右那么你还是再继续调速度吧。。。另外问下你是不是在拍太阳？？？那可不好，
如何使用佳能500D拍摄各种照片 编辑: 天马行空发表日期: 2011-04-19 15:02 写个简略单纯教程，也同时请认识单反认识500D的町友给予指正，因为我拍照也是瞎拍，所以有时说的难免有纰缪或不科学的处所，大家发现问题一定要及时给予究正。 买来新相机无论是老手还是新手都应当仔细阅读一遍说明书，然后再拿起相机亲自操作拍摄几天，然后回过头来再看一遍说明书，然后再拍几天，再看一遍说明书，一般至少要看三遍吧。。。横竖我是第五次看说明书的时候还能从其中学到新东西。因为说明书里有些内部实质意义在你没有实际拍摄经验的情况下是看不懂的，所以说明书一定要多看，说明书是最佳的教材。 全数内部实质意义如下： 1 怎样不错曝光 2 光圈快门速度ISO之间的瓜葛 3 怎样拍背景虚化的照片 4 镜头 焦距的选择 5 怎样拍夜景 6 被摄体不在画面中间时怎样对焦 7 如安在黑天室外拍人物像 8 怎样拍摄动态汽车照片 9 怎样拍摄鬼影照片 10 怎样拍摄剪影照片 11 怎样拍摄烟花 12 办婚仪式上怎样拍照 13 桌球室拍照 14 怎样拍显个，怎样拍显瘦，哪个角度最标致。 注意：和我学可学不可大师，极可能还学些凑弊端往后欠好改，不过按我说的做能在短时间内在外行面前假充假充妙手 现在开始按我的套路来了。。。 1 怎样不错曝光 先把相机拍摄模式调到M档，也就是全手动档。（如果不明白怎样调到M档就看一下说明书，往后就再也不说了然，看不懂的处所就翻说明书），有人说M档不实用操作贫苦，新手不易掌握，但我感觉M档更适合新手进修，因为M档需要手动调圈和快门更容易让新手迅速的掌握光圈和快门的瓜葛以及对成像的影响。横竖我拍照只要么是很紧急的情况下拍摄都是M档的。 要想用M档拍照先要知道怎样调光圈和快门，从液晶屏上看横排最上方有3个数据，最左面的是快门速度，中间是光劝巨细，右面是ISO感光度值。没事了在光线不是很暗的情况下我们拍照都要把ISO值设成一百，如果光线很暗又不想研发光灯的话可以适当调高。然后光圈和速度怎么调这个是不固定的，要看你所在的环境光线和你要拍摄的物体。别急，一步步来…… 说一下M档下怎样调快门速度光圈和ISO，直接转动节制转盘就可以调解快门速度，按住液晶屏右上角那Av按钮的同时转动转盘就是调光圈巨细，节制转盘边上有个ISO按钮，按一下然后用机背的标的目的键选择ISO，选好后按标的目的键中间的SET按钮确定。 知道怎么调往后我们就要热战演练了，先找个布娃娃或别的差不多的东西或是家里的任意同样安排，先拿它练起。这时候你要拿媳妇练预计容易给你媳妇练出拍照惧怕症~没事了理况下1
篇一：佳能500d pas 佳能500d pas-c数码单反 详细参数，如下： 主要性能发布日期机身特性操作方式传感器类型 2009年03月 aps-c规格数码单反全手动支持 cmos 长宽比：3:2 传感器件尺寸 22.3×14.9mm 传感器描述有效像素数影像处理器最高分辨率 除尘功能：自动，手动，添加除尘数据 1510万 digic 4 4752×3168 大：约1510万像素（4752×3168）中：约800万像素（3456×2304） 图像分辨率 小：约370万像素（2352×1568） raw：约1510万像素（4752×3168）全高清(1080p) 高清摄像镜头特点 佳能ef系列镜头（包括ef-s系列镜头）（35mm换算焦距约为镜头焦距的1.6 镜头说明 倍） 类型：ttl辅助影像重合，相位检测自动对焦点：9个自动对焦点 测光范围：ev -0.5-18（23℃，73℉，iso 100） 对焦方式 对焦模式：单次自动对焦，人工智能伺服自动对焦，人工智能自动对焦，手动对焦（mf）自动对焦辅助光：由内置闪光灯发出的短促连续闪光 9点变焦环 3英寸 类型：tft彩色液晶监视器（可分7级调节亮度）监视器尺寸和点数：3英寸，约92万点（vga） 对焦点数变焦方式显示功能显示屏尺寸 像素及类型 视野率：约100% 界面语言：25种（含简体中文） 可旋转液晶屏否触摸屏是否双屏取景器 不支持否光学 类型：眼平五面镜 视野率：垂直/水平方向约95% 放大倍率：约0.87倍（-1m-1，使用50mm镜头对无限远处对焦）眼点：约19mm（自目镜透镜中央起-1m-1） 取景器描述 内置屈光度调节：-3.0-+1.0m-1（dpt）对焦屏：固定式，精确磨砂反光镜：快回型景深预视 快门性能快门类型快门速度闪光灯闪光类型外接闪光灯(热靴) 电子控制纵走式焦平面快门 1/60-1/4000秒（全自动模式），闪光同步速度1/200秒 30-1/4000，b快门（总快门速度范围。可用范围随拍摄模式各异。
Traditional Chinese - Number 68e Nutrition Series - January 2017 食物來源中的鈣與維生素D Food Sources of Calcium and Vitamin D 我為何需要鈣與維生素D？ 鈣是強健骨骼的重要元素之一。人的一生中，需要不斷攝入含鈣食物，因為我們的骨骼始終在經歷損傷和重建的過程。 攝入含鈣的食物，有助於防止骨骼變得脆弱，從而預防骨折或損傷。維生素D能幫助人體吸收鈣質， 在人體健康的其他方面（例如免疫、心血管健康和癌症預防等）也可能發揮著作用。目前尚不確定維生素D具體能在多少方面發揮作用。 我需要多少鈣和維生素D？ 哪些食物中含有鈣？ 許多不同食物中都含有少量的鈣。富含鈣質的食物包括奶類、酸奶和添加了鈣質的飲品，例如強化橙汁和部分以植物為原料的飲品。奶酪也是鈣質的優質來源。食物包裝上的營養成分標籤會列出其中所含鈣質所佔人體每日所需的百分比（DV%）。鈣含量佔人體每日所需15%或以上的食物，可以認為是富含鈣質的。如果食物包裝上沒有營養成分標籤，請您查閱“鈣質的食物來源”表，瞭解相關內容。 哪些食物含維生素D？ 僅有很少的食物中含有維生素D。維生素D含量較高的食物包括魚、肝臟和蛋黃。添加維生素D的食物和飲料是維生素D的優質來源。牛 奶中一定會添加維生素D，強化型橙汁、人造黃油和許 多以植物為主要原料的飲品中也有添加。對於其他食物，請在食品包裝上尋找“人工添加維生素D”（fortified with vitamin D）或“強化維生素D”（enriched with vitamin D）的字樣。人工添加和強化類的食物，指的是在其中添加 了某些特定的營養元素。 我应该服用維生素D補充劑嗎？ 請參考“維生素D的食物來源”一表，瞭解您一般需要攝 入多少維生素D。您可能會發現，僅通過食物無法攝取 到足夠的維生素D。您可以服用單獨的維生素D補充劑，或含有維生素D的複合維生素。 以下兩類人士，每日應服用400IU的維生素D補充劑： ?正在母乳餵哺，或在母乳餵哺的同時也喝嬰兒配方奶的嬰兒和幼童；以及 ?50歲以上的成人。 我应该服用鈣質補充劑嗎？ 參考“鈣質的食物來源”一表中的各類食物，選擇攝入種 類豐富的食物來滿足您的鈣質所需。如果您覺得僅通過 食物無法攝取到足夠的鈣質，請向您的醫護人員諮詢適 合您的鈣質補充劑的種類及用量。 食物來源中的維生素D
佳能数码相机维修站教你如何使用佳能500D 点击：1462 时间：2009-07-28 佳能500D的保养 本相机是精密仪器。请勿将其摔落或使其受到物理撞击。本相机不是防水相机，不能在水下使用。如果相机不慎落入水中，请立即向 附近的佳能维修中心咨询。请用干布拭去水珠。如果相机曾暴露在含盐分的空气中，请用拧干的湿布擦拭。 请勿将本相机靠近具有强磁场的物体，如磁铁或电动机。另外也要避免将相机靠近发出较强无线电波的物体，如天线。强磁场可能引起相机故障或破坏图像数据。 请勿将本相机放在温度过高的地方，如处于阳光直射的汽车内。高温可能导致相机故障。相机内有精密电子线路。请勿自行拆卸相机。请使用气吹吹走镜头、取景器、反光镜和对焦屏上的灰尘。请勿使用含有有 机溶剂的清洁剂清洁机身和镜头。对于顽固污渍，请将相机送到附近的佳能维修中心处理。请勿用手指接触相机的电子触点，以免触点受到腐蚀。腐蚀的触点可能导致 相机故障。如果相机突然从低温处进入温暖的房间，可能造成相机表面和其内部零件结露。为防止结露，请先将相机放入密封的塑料袋中，然后等其温度逐步升高 后再从袋中取出。如果相机出现结露，请勿使用，以免损坏相机。如果发生这种情况，请从相机上卸下镜头，取出存储卡和电池，等到结露蒸发后再使用相机。 如果相机长时间不使用，请取出电池并将相机放置在通风良好的干燥阴凉处。存放期间请隔一段时间按动几次快门，以确认相机是否能正常工作。 避免将相机存放在暗房、实验室等有腐蚀性化学物品的地方。如果长时间未使用相机，拍摄前请测试所有功能。如果相机长时间未使用或 即将进行重要拍摄活动，请将相机送交经销商检测或自行进行检测，并确认相机工作正常。操作注意事项 液晶监视器： 虽然液晶监视器是采用高精密技术制造的，超过99.99％的像素为有效像素，但是剩余0.01％或更少的像素中可能存在若干坏点。坏点总是显示为黑色或 红色等颜色，这并不是故障。坏点并不影响图像记录效果。如果液晶监视器长时间保持开启状态，屏幕可能会出现异常，从屏幕上可以 看到所显示图像的残像。但是，这种情况只是暂时的，不使用相机几天以后，它便会消失。在低温或高温条件下，液晶监视器可能会显示较慢或看起来有些黑。它会在 室温下恢复正常。 存储卡 为保护存储卡和记录的数据不受损坏，请注意下列各项：请勿将存储卡掉落、弯折或弄湿。请勿对其施加过大外力、使其受到撞击或震动。 请勿在有任何强磁场的物体附近存放或使用存储卡，如电视机、扬声器或磁铁。另外要避免易于产生静电的场所。请勿将存储卡置于阳光下曝晒或靠近热源。 将存储卡存放在盒中。请勿将存储卡存放在高温、多尘或潮湿的环境中。 镜头 将镜头从机身卸下后，请装上镜头盖或将镜头按头朝下方式触点 竖直放置，避免刮擦镜头表面和电子触点。