CN113923444B - Zoom lens quality evaluation method and device - Google Patents
Zoom lens quality evaluation method and device Download PDFInfo
- Publication number
- CN113923444B CN113923444B CN202111170279.4A CN202111170279A CN113923444B CN 113923444 B CN113923444 B CN 113923444B CN 202111170279 A CN202111170279 A CN 202111170279A CN 113923444 B CN113923444 B CN 113923444B
- Authority
- CN
- China
- Prior art keywords
- acquiring
- preset
- feature
- value
- evaluation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000013441 quality evaluation Methods 0.000 title claims abstract description 38
- 238000011156 evaluation Methods 0.000 claims abstract description 138
- 238000003384 imaging method Methods 0.000 claims abstract description 54
- 238000006243 chemical reaction Methods 0.000 claims abstract description 9
- 238000012795 verification Methods 0.000 claims description 78
- 238000004364 calculation method Methods 0.000 claims description 36
- 238000013210 evaluation model Methods 0.000 claims description 26
- 230000008569 process Effects 0.000 claims description 20
- 230000001502 supplementing effect Effects 0.000 claims description 18
- 238000012216 screening Methods 0.000 claims description 16
- 230000004069 differentiation Effects 0.000 claims description 6
- 230000003287 optical effect Effects 0.000 abstract description 10
- 238000004519 manufacturing process Methods 0.000 abstract description 8
- 230000000875 corresponding effect Effects 0.000 description 130
- 239000000463 material Substances 0.000 description 14
- 230000009286 beneficial effect Effects 0.000 description 10
- 230000002596 correlated effect Effects 0.000 description 10
- 238000010801 machine learning Methods 0.000 description 10
- 238000012163 sequencing technique Methods 0.000 description 8
- FFBHFFJDDLITSX-UHFFFAOYSA-N benzyl N-[2-hydroxy-4-(3-oxomorpholin-4-yl)phenyl]carbamate Chemical compound OC1=C(NC(=O)OCC2=CC=CC=C2)C=CC(=C1)N1CCOCC1=O FFBHFFJDDLITSX-UHFFFAOYSA-N 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000009193 crawling Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000005669 field effect Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N17/00—Diagnosis, testing or measuring for television systems or their details
- H04N17/002—Diagnosis, testing or measuring for television systems or their details for television cameras
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Studio Devices (AREA)
- Eyeglasses (AREA)
Abstract
The invention provides a zoom lens quality evaluation method and device, wherein the method comprises the following steps: acquiring imaging data of a zoom lens, and performing photoelectric conversion on the imaging data to obtain data to be evaluated; and carrying out quality evaluation on the zoom lens based on the data to be evaluated. According to the method and the device for evaluating the quality of the zoom lens, the quality of the zoom lens is evaluated based on the data to be evaluated, so that workers can be helped to rapidly evaluate the imaging quality of the lens in production, the imaging quality is improved, the reject ratio is reduced, and the enterprise cost is reduced by utilizing the evaluation result to carry out optical adjustment on the zoom lens.
Description
Technical Field
The invention relates to the technical field of lens evaluation, in particular to a zoom lens quality evaluation method and device.
Background
At present, along with the development of the optical industry, higher requirements are made on the definition, object image similarity and deformation degree of an image formed by an optical system, so that the development of a zoom lens is promoted, the precision requirements of lenses and related lens frames are required to be improved to meet the requirement of zooming, but the precision improvement is not only limited by the existing processing technology, but also the production cost of enterprises is increased.
Therefore, there is a need for a method for evaluating quality of a zoom lens, which rapidly evaluates imaging quality of the lens in production, and uses the evaluation result to optically adjust the zoom lens, thereby improving imaging quality, reducing reject ratio, and reducing enterprise cost.
Disclosure of Invention
The invention aims to provide a quality evaluation method and device for a zoom lens, which are used for evaluating the quality of the zoom lens based on data to be evaluated, and can help staff to rapidly evaluate the imaging quality of the lens in production, and the evaluation result is used for carrying out optical adjustment on the zoom lens, so that the imaging quality is improved, the reject ratio is reduced, and the enterprise cost is reduced.
The zoom lens quality evaluation method provided by the embodiment of the invention comprises the following steps:
Acquiring imaging data of a zoom lens, and performing photoelectric conversion on the imaging data to obtain data to be evaluated;
And carrying out quality evaluation on the zoom lens based on the data to be evaluated.
Preferably, acquiring imaging data of the zoom lens includes:
Providing a monochromatic light source;
And transmitting the light beams emitted by the monochromatic light sources to pass through the differentiation plate and the zoom lens in sequence, and imaging on a high-precision linear array CCD to obtain imaging data.
Preferably, the quality evaluation of the zoom lens based on the data to be evaluated includes:
Acquiring a preset evaluation target set, wherein the evaluation target set comprises: a plurality of first evaluation targets;
Acquiring attribute information of a zoom lens, and screening a second evaluation target to be evaluated from the first evaluation targets based on the attribute information;
Extracting target data corresponding to a second evaluation target from the data to be evaluated;
Determining an evaluation model corresponding to a second evaluation target based on a preset evaluation target-evaluation model library;
And inputting the target data into a corresponding evaluation model to obtain an evaluation result.
Preferably, the step of screening out a second evaluation target to be evaluated from the first evaluation targets based on the attribute information includes:
extracting a plurality of first attribute items in the attribute information;
acquiring the attribute type of the first attribute item, and determining an important value corresponding to the attribute type based on a preset attribute type-important value library;
if the importance value is greater than or equal to a preset importance value threshold value, the corresponding first attribute item is used as a second attribute item;
acquiring a preset negative event generation model, inputting a second attribute item into the negative event generation model, and acquiring at least one negative event and a first serious value corresponding to the negative event;
acquiring a preset capture strategy set, wherein the capture strategy set comprises: a plurality of capture strategies;
Determining at least one capturing object corresponding to the capturing strategy based on a preset capturing strategy-capturing object library;
attempting to capture at least one first lens quality event in the capture object based on the capture policy;
if the capturing is successful, acquiring a capturing process for capturing the first lens quality event;
carrying out flow splitting on the capturing flow to obtain a plurality of flows;
Extracting at least one first captured scene in the process;
Acquiring the credibility of a first capturing scene;
if the credibility is smaller than or equal to a preset credibility threshold, taking the corresponding first captured scene as a second captured scene;
attempting to acquire an association relationship between the second captured scene and the captured object;
If the acquisition fails, rejecting a corresponding first lens quality event;
If the acquisition is successful, analyzing the association relation to acquire a relation value;
If the relation value is smaller than or equal to a preset relation value threshold value, eliminating the corresponding first lens quality event;
When all the first lens quality events needing to be removed in the first lens quality events are removed, taking the remaining first lens quality events as second lens quality events;
Acquiring a preset receiving node set, wherein the receiving node set comprises: a plurality of receiving nodes;
On the basis of a preset query mode, the receiving node is queried at regular time, and at least one third shot quality event replied by the queried receiving node and at least one sender corresponding to the third shot quality event are acquired;
acquiring a sending record of a sender;
extracting a plurality of first record items in the transmission record;
Establishing a time axis, and setting the first record item on a corresponding time node on the time axis based on the generation time of the first record item;
Extracting features of the first record item to obtain a plurality of first features;
Acquiring a preset risk feature library, matching the first feature with a second feature in the risk feature library, taking the matched second feature as a third feature if the matching is met, and taking a corresponding first record item as a second record item;
Based on a preset feature-supplementing direction library, attempting to determine at least one supplementing direction corresponding to the third feature;
if the determination fails, randomly combining the first features obtained by extracting the features of the second record item to obtain a plurality of first combined features;
acquiring a preset malicious feature library, matching the first combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
If the determination is successful, selecting at least one first record item in a range preset in the supplementing direction of the second record item on the time axis, and taking the first record item as a third record item;
Extracting the characteristics of the third record item to obtain a plurality of fifth characteristics;
Randomly combining the first feature and the fifth feature obtained by extracting the features of the second record item to obtain a plurality of second combined features;
matching the second combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
When all the third lens quality events needing to be removed in the third lens quality events are removed, taking the rest third lens quality events as fourth lens quality events;
Acquiring a preset event determination model, determining whether a negative event is contained in a second lens quality event and a fourth lens quality event or not by the event determination model, and if so, outputting a first serious value corresponding to the contained negative event and taking the first serious value as a second serious value;
Summarizing the second serious value to obtain a sorting value;
Sequencing the first evaluation targets corresponding to the first attribute items according to the sizes of the corresponding sequencing values to obtain an evaluation target sequence;
and selecting first n first evaluation targets in the evaluation target sequence as second evaluation targets, and finishing screening.
Preferably, the zoom lens quality evaluation method further includes:
Expanding a risk feature library;
The method for expanding the risk feature library comprises the following steps:
obtaining a preset extended node set, wherein the extended node set comprises: a plurality of first expansion nodes;
Acquiring the guarantee information of the first expansion node, wherein the guarantee information comprises the following components: a second extended node for guaranteeing the first extended node and a first guaranteeing value corresponding to the second extended node;
Determining a first expansion node in the second expansion nodes, taking the first expansion node as a third expansion node, and taking a first holding value corresponding to the third expansion node as a second holding value;
Taking a second expansion node except a third expansion node in the second expansion nodes as a fourth expansion node, and taking a first held value corresponding to the fourth expansion node as a third held value;
Acquiring a preset first calculation model, inputting the second guaranteed value and the third guaranteed value into the first calculation model, and acquiring a first score;
if the first score is greater than or equal to a preset first score threshold, the corresponding first expansion node is used as a fifth expansion node;
acquiring at least one first risk feature through a fifth expansion node;
Acquiring a preset forward verification model, inputting a first risk feature into the forward verification model, and acquiring at least one forward verification value;
Acquiring a preset reverse verification model, inputting the first risk characteristic into the reverse verification model, and acquiring at least one reverse verification value;
Acquiring a preset second calculation model, and inputting the forward verification value and the reverse verification value into the second calculation model to acquire a second score;
If the second score is greater than or equal to a preset second score threshold, the corresponding first risk feature is used as a second risk feature;
storing the second risk features into a risk feature library;
And after the second risk features which are required to be stored in the risk feature library are stored, completing expansion.
The device for evaluating the quality of the zoom lens provided by the embodiment of the invention comprises the following components:
the acquisition module is used for acquiring imaging data of the zoom lens, and performing photoelectric conversion on the imaging data to acquire data to be evaluated;
and the evaluation module is used for evaluating the quality of the zoom lens based on the data to be evaluated.
Preferably, the acquisition module performs the following operations:
Providing a monochromatic light source;
And transmitting the light beams emitted by the monochromatic light sources to pass through the differentiation plate and the zoom lens in sequence, and imaging on a high-precision linear array CCD to obtain imaging data.
Preferably, the evaluation module performs the following operations:
Acquiring a preset evaluation target set, wherein the evaluation target set comprises: a plurality of first evaluation targets;
Acquiring attribute information of a zoom lens, and screening a second evaluation target to be evaluated from the first evaluation targets based on the attribute information;
Extracting target data corresponding to a second evaluation target from the data to be evaluated;
Determining an evaluation model corresponding to a second evaluation target based on a preset evaluation target-evaluation model library;
And inputting the target data into a corresponding evaluation model to obtain an evaluation result.
Preferably, the evaluation module performs the following operations:
extracting a plurality of first attribute items in the attribute information;
acquiring the attribute type of the first attribute item, and determining an important value corresponding to the attribute type based on a preset attribute type-important value library;
if the importance value is greater than or equal to a preset importance value threshold value, the corresponding first attribute item is used as a second attribute item;
acquiring a preset negative event generation model, inputting a second attribute item into the negative event generation model, and acquiring at least one negative event and a first serious value corresponding to the negative event;
acquiring a preset capture strategy set, wherein the capture strategy set comprises: a plurality of capture strategies;
Determining at least one capturing object corresponding to the capturing strategy based on a preset capturing strategy-capturing object library;
attempting to capture at least one first lens quality event in the capture object based on the capture policy;
if the capturing is successful, acquiring a capturing process for capturing the first lens quality event;
carrying out flow splitting on the capturing flow to obtain a plurality of flows;
Extracting at least one first captured scene in the process;
Acquiring the credibility of a first capturing scene;
if the credibility is smaller than or equal to a preset credibility threshold, taking the corresponding first captured scene as a second captured scene;
attempting to acquire an association relationship between the second captured scene and the captured object;
If the acquisition fails, rejecting a corresponding first lens quality event;
If the acquisition is successful, analyzing the association relation to acquire a relation value;
If the relation value is smaller than or equal to a preset relation value threshold value, eliminating the corresponding first lens quality event;
When all the first lens quality events needing to be removed in the first lens quality events are removed, taking the remaining first lens quality events as second lens quality events;
Acquiring a preset receiving node set, wherein the receiving node set comprises: a plurality of receiving nodes;
On the basis of a preset query mode, the receiving node is queried at regular time, and at least one third shot quality event replied by the queried receiving node and at least one sender corresponding to the third shot quality event are acquired;
acquiring a sending record of a sender;
extracting a plurality of first record items in the transmission record;
Establishing a time axis, and setting the first record item on a corresponding time node on the time axis based on the generation time of the first record item;
Extracting features of the first record item to obtain a plurality of first features;
Acquiring a preset risk feature library, matching the first feature with a second feature in the risk feature library, taking the matched second feature as a third feature if the matching is met, and taking a corresponding first record item as a second record item;
Based on a preset feature-supplementing direction library, attempting to determine at least one supplementing direction corresponding to the third feature;
if the determination fails, randomly combining the first features obtained by extracting the features of the second record item to obtain a plurality of first combined features;
acquiring a preset malicious feature library, matching the first combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
If the determination is successful, selecting at least one first record item in a range preset in the supplementing direction of the second record item on the time axis, and taking the first record item as a third record item;
Extracting the characteristics of the third record item to obtain a plurality of fifth characteristics;
Randomly combining the first feature and the fifth feature obtained by extracting the features of the second record item to obtain a plurality of second combined features;
matching the second combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
When all the third lens quality events needing to be removed in the third lens quality events are removed, taking the rest third lens quality events as fourth lens quality events;
Acquiring a preset event determination model, determining whether a negative event is contained in a second lens quality event and a fourth lens quality event or not by the event determination model, and if so, outputting a first serious value corresponding to the contained negative event and taking the first serious value as a second serious value;
Summarizing the second serious value to obtain a sorting value;
Sequencing the first evaluation targets corresponding to the first attribute items according to the sizes of the corresponding sequencing values to obtain an evaluation target sequence;
and selecting first n first evaluation targets in the evaluation target sequence as second evaluation targets, and finishing screening.
Preferably, the zoom lens quality evaluation apparatus further comprises:
the expansion module is used for expanding the risk feature library;
The expansion module performs the following operations:
obtaining a preset extended node set, wherein the extended node set comprises: a plurality of first expansion nodes;
Acquiring the guarantee information of the first expansion node, wherein the guarantee information comprises the following components: a second extended node for guaranteeing the first extended node and a first guaranteeing value corresponding to the second extended node;
Determining a first expansion node in the second expansion nodes, taking the first expansion node as a third expansion node, and taking a first holding value corresponding to the third expansion node as a second holding value;
Taking a second expansion node except a third expansion node in the second expansion nodes as a fourth expansion node, and taking a first held value corresponding to the fourth expansion node as a third held value;
Acquiring a preset first calculation model, inputting the second guaranteed value and the third guaranteed value into the first calculation model, and acquiring a first score;
if the first score is greater than or equal to a preset first score threshold, the corresponding first expansion node is used as a fifth expansion node;
acquiring at least one first risk feature through a fifth expansion node;
Acquiring a preset forward verification model, inputting a first risk feature into the forward verification model, and acquiring at least one forward verification value;
Acquiring a preset reverse verification model, inputting the first risk characteristic into the reverse verification model, and acquiring at least one reverse verification value;
Acquiring a preset second calculation model, and inputting the forward verification value and the reverse verification value into the second calculation model to acquire a second score;
If the second score is greater than or equal to a preset second score threshold, the corresponding first risk feature is used as a second risk feature;
storing the second risk features into a risk feature library;
And after the second risk features which are required to be stored in the risk feature library are stored, completing expansion.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
FIG. 1 is a flowchart of a zoom lens quality evaluation method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a specific application of quality evaluation of a zoom lens according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a zoom lens quality evaluation apparatus according to an embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings, it being understood that the preferred embodiments described herein are for illustration and explanation of the present invention only, and are not intended to limit the present invention.
The embodiment of the invention provides a zoom lens quality evaluation method, as shown in fig. 1, comprising the following steps:
step S1: acquiring imaging data of a zoom lens, and performing photoelectric conversion on the imaging data to obtain data to be evaluated;
Step S2: and carrying out quality evaluation on the zoom lens based on the data to be evaluated.
The working principle and the beneficial effects of the technical scheme are as follows:
acquiring imaging data imaged by the zoom lens (the imaging is performed on an imaging device by using a light source to emit a light beam to pass through the lens, the technology belongs to the prior art and is not described in detail), and because the imaging data belongs to optical information, photoelectric conversion is required to be performed to convert the optical information into data to be evaluated which belongs to electrical information; based on the data to be evaluated, performing quality evaluation (such as evaluation of imaging face value, contrast, focal length, depth of field, field curvature and the like) on the zoom lens;
According to the embodiment of the invention, the quality evaluation is carried out on the zoom lens based on the data to be evaluated, so that a worker can be helped to rapidly evaluate the imaging quality of the lens in production, the imaging quality is improved, the reject ratio is reduced, and the enterprise cost is reduced by utilizing the evaluation result to carry out optical adjustment on the zoom lens.
The embodiment of the invention provides a zoom lens quality evaluation method, which comprises the following steps of:
Providing a monochromatic light source;
And transmitting the light beams emitted by the monochromatic light sources to pass through the differentiation plate and the zoom lens in sequence, and imaging on a high-precision linear array CCD to obtain imaging data.
The working principle and the beneficial effects of the technical scheme are as follows:
As shown in fig. 2, a monochromatic light source is provided (for example, a light source machine is used); the light beam emitted by the monochromatic light source is transmitted through the dividing plate (the dividing plate has the function of superposing a cross wire or concentric ring pattern on an object to be imaged, the pattern can be used as a position reference and can be aligned with the object to be imaged, the cross dividing plate or concentric ring dividing plate can be used) and the zoom lens, imaging is carried out on the high-precision linear array CCD (Charge Coupled Device ) and the acquisition of imaging data is completed.
The embodiment of the invention provides a quality evaluation method of a zoom lens, which is used for evaluating the quality of the zoom lens based on data to be evaluated and comprises the following steps:
Acquiring a preset evaluation target set, wherein the evaluation target set comprises: a plurality of first evaluation targets;
Acquiring attribute information of a zoom lens, and screening a second evaluation target to be evaluated from the first evaluation targets based on the attribute information;
Extracting target data corresponding to a second evaluation target from the data to be evaluated;
Determining an evaluation model corresponding to a second evaluation target based on a preset evaluation target-evaluation model library;
And inputting the target data into a corresponding evaluation model to obtain an evaluation result.
The working principle and the beneficial effects of the technical scheme are as follows:
The first evaluation target is specifically: for example, evaluating contrast, evaluating depth of field effect, and the like; the preset evaluation target-evaluation model library specifically comprises the following components: a database containing evaluation models corresponding to different evaluation targets, for example: the evaluation target is evaluation contrast, and the corresponding evaluation model is a model generated after a large number of records of manual evaluation contrast are learned by using a machine learning algorithm and is used for evaluating the contrast; the attribute information of the zoom lens specifically includes: for example, material provider information of the lens, production process of the lens, manufacturer, etc.;
After the second evaluation target is determined, target data (for example, contrast data) corresponding to the evaluation target in the data to be evaluated is determined, and the evaluation is performed by using a corresponding evaluation model.
The embodiment of the invention provides a zoom lens quality evaluation method, which screens out a second evaluation target to be evaluated from first evaluation targets based on attribute information, and comprises the following steps:
extracting a plurality of first attribute items in the attribute information;
acquiring the attribute type of the first attribute item, and determining an important value corresponding to the attribute type based on a preset attribute type-important value library;
if the importance value is greater than or equal to a preset importance value threshold value, the corresponding first attribute item is used as a second attribute item;
acquiring a preset negative event generation model, inputting a second attribute item into the negative event generation model, and acquiring at least one negative event and a first serious value corresponding to the negative event;
acquiring a preset capture strategy set, wherein the capture strategy set comprises: a plurality of capture strategies;
Determining at least one capturing object corresponding to the capturing strategy based on a preset capturing strategy-capturing object library;
attempting to capture at least one first lens quality event in the capture object based on the capture policy;
if the capturing is successful, acquiring a capturing process for capturing the first lens quality event;
carrying out flow splitting on the capturing flow to obtain a plurality of flows;
Extracting at least one first captured scene in the process;
Acquiring the credibility of a first capturing scene;
if the credibility is smaller than or equal to a preset credibility threshold, taking the corresponding first captured scene as a second captured scene;
attempting to acquire an association relationship between the second captured scene and the captured object;
If the acquisition fails, rejecting a corresponding first lens quality event;
If the acquisition is successful, analyzing the association relation to acquire a relation value;
If the relation value is smaller than or equal to a preset relation value threshold value, eliminating the corresponding first lens quality event;
When all the first lens quality events needing to be removed in the first lens quality events are removed, taking the remaining first lens quality events as second lens quality events;
Acquiring a preset receiving node set, wherein the receiving node set comprises: a plurality of receiving nodes;
On the basis of a preset query mode, the receiving node is queried at regular time, and at least one third shot quality event replied by the queried receiving node and at least one sender corresponding to the third shot quality event are acquired;
acquiring a sending record of a sender;
extracting a plurality of first record items in the transmission record;
Establishing a time axis, and setting the first record item on a corresponding time node on the time axis based on the generation time of the first record item;
Extracting features of the first record item to obtain a plurality of first features;
Acquiring a preset risk feature library, matching the first feature with a second feature in the risk feature library, taking the matched second feature as a third feature if the matching is met, and taking a corresponding first record item as a second record item;
Based on a preset feature-supplementing direction library, attempting to determine at least one supplementing direction corresponding to the third feature;
if the determination fails, randomly combining the first features obtained by extracting the features of the second record item to obtain a plurality of first combined features;
acquiring a preset malicious feature library, matching the first combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
If the determination is successful, selecting at least one first record item in a range preset in the supplementing direction of the second record item on the time axis, and taking the first record item as a third record item;
Extracting the characteristics of the third record item to obtain a plurality of fifth characteristics;
Randomly combining the first feature and the fifth feature obtained by extracting the features of the second record item to obtain a plurality of second combined features;
matching the second combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
When all the third lens quality events needing to be removed in the third lens quality events are removed, taking the rest third lens quality events as fourth lens quality events;
Acquiring a preset event determination model, determining whether a negative event is contained in a second lens quality event and a fourth lens quality event or not by the event determination model, and if so, outputting a first serious value corresponding to the contained negative event and taking the first serious value as a second serious value;
Summarizing the second serious value to obtain a sorting value;
Sequencing the first evaluation targets corresponding to the first attribute items according to the sizes of the corresponding sequencing values to obtain an evaluation target sequence;
and selecting first n first evaluation targets in the evaluation target sequence as second evaluation targets, and finishing screening.
The working principle and the beneficial effects of the technical scheme are as follows:
The preset attribute type-important value library specifically comprises the following steps: a database containing importance values corresponding to different attribute types, for example: the attribute type is a lens material provider, the corresponding importance value is 95, and the importance value represents the influence degree of the attribute type on the lens quality; the preset threshold value of the important value is specifically: for example, 75; the preset negative event generation model specifically comprises the following steps: a model generated after learning a number of manually made records of negative events using a machine learning algorithm, the model may generate negative events based on attribute terms, such as: the attribute item is a lens material manufacturer, the negative event is that the lens material produced by the manufacturer causes unqualified lens quality and the like, a first serious value of the negative event is output, and the larger the first serious value is, the more serious the corresponding negative event is; the capture strategy is specifically: for example, web crawling; the preset capturing strategy-capturing object library specifically comprises the following steps: data containing capture objects corresponding to different capture strategies, such as: the step strategy is that a webpage is crawled, and the capturing objects are a plurality of lens forum websites; the preset credibility threshold is specifically: for example 80; the preset relation value threshold value is specifically: for example, 90; the receiving node specifically comprises: a network node to which a peer can send data, i.e. complete data reception; the preset query mode specifically comprises the following steps: for example, if there is newly received data, if so, please reply; the preset risk feature library specifically comprises the following steps: a database containing a plurality of risk features, the risk features being features that may lead to malicious events, such as: the user wants to issue false data, suddenly changing identity information (risk feature); the preset malicious feature library specifically comprises the following steps: a database containing a plurality of malicious features, such as: changing identity information, and the released data is detected as false data; the preset range is specifically as follows: a range on the time axis corresponding to a certain length of time (e.g., 2 days); the preset event determination model specifically comprises the following steps: a model generated after the record of manually judging whether the event contains a certain event is learned by utilizing a machine learning algorithm, wherein the model can determine whether the event contains a certain event;
When screening the second evaluation target to be evaluated, it may be determined based on the attribute information of the zoom lens, for example: the manufacturers for producing the material of the zoom lens have the problem that the contrast of imaging of the lens is poor due to the material problem, and then the contrast evaluation is required; how to acquire problem events, however, needs to be resolved; the application extracts a first attribute item (such as a material provider) in attribute information, screens out a second attribute item which has influence on a quality result based on an important value, generates a negative event corresponding to the second attribute item based on a negative event generation model, and determines whether the negative event occurs or not; attempting to capture a shot quality event (for example, a quality evaluation article published by a user of a forum website) in a capture object, if capturing is successful, determining a plurality of processes in a capture process (for example, process 1, capturing a comment of "quality is not feasible" published by the user in a factory discussion area, process 2, capturing a homepage of the user to publish a specific article), wherein a first capture scene corresponding to the process is, for example: the process 1 corresponds to a web page of a discussion area of a manufacturer, and the process 2 corresponds to a web page of a user homepage; acquiring the credibility (webpage credibility) of the first captured scene, and taking the corresponding first captured scene as a second captured scene if the credibility is smaller than a credibility threshold; attempting to acquire the association relation between the second captured scene and the captured object (for example, the second captured scene belongs to a webpage of the captured object, the captured object guarantees the webpage to form a guarantee relation), if the acquisition fails, indicating that the second captured scene is not trusted (for example, from other forums), if the acquisition is successful, analyzing the relation (guarantee relation), acquiring a relation value, wherein the larger the relation value is, the larger the guarantee degree of the captured object guaranteeing the relation value is, and the reliability is high; some cooperative forums do not allow my to capture for privacy security, may set receiving nodes, and the counterpart may send to my; but when a third shot quality event is received, the corresponding sender needs to be verified; acquiring a sending record (record of data sent to a plurality of receiving nodes historically, etc.) of a sender, extracting a plurality of first record items, matching a first feature with a second feature, determining whether risk features possibly causing malicious events occur, if so, determining a supplementing direction, for example: the third feature is that the user changes the identity information suddenly, possibly uploading false data later, and the supplementing direction is later; if the supplementing direction cannot be determined, indicating that a malicious event exists in the second record item, respectively combining the first combination feature and the second combination feature based on the two conditions, and matching the first combination feature and the second combination feature with the fourth feature to determine whether the malicious event occurs; after eliminating the lens quality event to be eliminated, determining whether the rest lens quality event contains a corresponding negative event, wherein the larger the content is, the more the evaluation target corresponding to the first attribute item needs to be evaluated;
When the second evaluation target to be evaluated is selected from the first evaluation targets, the second evaluation targets can be determined based on the attribute information of the zoom lens, the setting is reasonable, the quality evaluation efficiency of the zoom lens is improved, and unnecessary workload is reduced; meanwhile, when acquiring the lens quality event, the unreliable lens quality event is removed, and the accuracy and safety of acquisition are ensured.
The embodiment of the invention provides a zoom lens quality evaluation method, which further comprises the following steps:
Expanding a risk feature library;
The method for expanding the risk feature library comprises the following steps:
obtaining a preset extended node set, wherein the extended node set comprises: a plurality of first expansion nodes;
Acquiring the guarantee information of the first expansion node, wherein the guarantee information comprises the following components: a second extended node for guaranteeing the first extended node and a first guaranteeing value corresponding to the second extended node;
Determining a first expansion node in the second expansion nodes, taking the first expansion node as a third expansion node, and taking a first holding value corresponding to the third expansion node as a second holding value;
Taking a second expansion node except a third expansion node in the second expansion nodes as a fourth expansion node, and taking a first held value corresponding to the fourth expansion node as a third held value;
Acquiring a preset first calculation model, inputting the second guaranteed value and the third guaranteed value into the first calculation model, and acquiring a first score;
if the first score is greater than or equal to a preset first score threshold, the corresponding first expansion node is used as a fifth expansion node;
acquiring at least one first risk feature through a fifth expansion node;
Acquiring a preset forward verification model, inputting a first risk feature into the forward verification model, and acquiring at least one forward verification value;
Acquiring a preset reverse verification model, inputting the first risk characteristic into the reverse verification model, and acquiring at least one reverse verification value;
Acquiring a preset second calculation model, and inputting the forward verification value and the reverse verification value into the second calculation model to acquire a second score;
If the second score is greater than or equal to a preset second score threshold, the corresponding first risk feature is used as a second risk feature;
storing the second risk features into a risk feature library;
And after the second risk features which are required to be stored in the risk feature library are stored, completing expansion.
The working principle and the beneficial effects of the technical scheme are as follows:
The first expansion node specifically comprises: the network node corresponds to a collector of the risk characteristics, and the risk characteristics collected by the collector can be obtained through the node; the preset first scoring threshold value specifically comprises the following steps: for example, 85; the preset forward verification model specifically comprises the following steps: the machine learning algorithm is utilized to learn a plurality of manual forward verification records (based on a plurality of malicious event data, whether the risk features can positively cause the occurrence of malicious events or not) to generate a model, and the larger the forward verification value is, the larger the occurrence probability of the malicious events caused by the risk features is; the preset reverse verification model specifically comprises the following steps: the method comprises the steps of utilizing a machine learning algorithm to learn a model generated after a large number of manual reverse verification records (based on a large number of malicious event data, whether risk features exist before a malicious event happens or not is verified) and enabling the probability of the risk features to be larger before the malicious event happens to be larger as a reverse verification value is larger; the preset second scoring threshold value specifically comprises the following steps: for example, 90;
when the first expansion node is collected Fang Duiying, other second expansion nodes are required to guarantee the first expansion node (guarantee is carried out, one side generates bad records, the other side also generates influences, for example, the guarantee value is reduced); calculating a first score based on the second guaranteed value and the third guaranteed value, wherein the higher the first score is, the higher the credibility of the corresponding first expansion node is; when the first risk feature is acquired, whether the risk feature is real or not needs to be verified in a forward and reverse direction, and a second risk feature meeting the condition is screened out to expand a risk feature library; the efficiency of the system for finding risk features can be further improved;
The preset first calculation model specifically comprises the following steps: a model with a built-in calculation formula, the calculation formula of which is as follows:
Wherein σ is the first score, A i is the i-th second held value, l is the total number of the second held values, B i is the i-th third held value, d is the total number of the third held values, ε 1 and ε 2 are preset weight values, Ρ is an intermediate variable;
In the formula, the second guarantee value and the third guarantee value are positively correlated with the first score, and because the third expansion node approved by the user corresponding to the second guarantee value guarantees the first expansion node to acquire the higher the guarantee approval degree, the l-d is positively correlated with the first score;
the preset second calculation model specifically comprises the following steps: a model with a built-in calculation formula, the calculation formula of which is as follows:
Wherein γ is the second score, e is a natural constant, α t is the t-th forward verification value, n 1 is the total number of the forward verification values, β t is the t-th reverse verification value, n 2 is the total number of the reverse verification values, ε is the sum of the first number of the forward verification values less than or equal to a preset first threshold value and the second number of the reverse verification values less than or equal to a preset second threshold value in the forward verification values;
In the formula, the positive verification value and the negative verification value are positively correlated with the second score, the first number and the second number are negatively correlated with the second score, and the sum of the first number and the second number is negatively correlated with the second score.
Through the formula, the first score and the second score are calculated rapidly, the expansion nodes and the risk features are convenient to screen, and the working efficiency of the system is improved to a great extent.
An embodiment of the present invention provides a zoom lens quality evaluation apparatus, as shown in fig. 3, including:
the acquisition module 1 is used for acquiring imaging data of the zoom lens, and performing photoelectric conversion on the imaging data to obtain data to be evaluated;
And the evaluation module 2 is used for evaluating the quality of the zoom lens based on the data to be evaluated.
The working principle and the beneficial effects of the technical scheme are as follows:
acquiring imaging data imaged by the zoom lens (the imaging is performed on an imaging device by using a light source to emit a light beam to pass through the lens, the technology belongs to the prior art and is not described in detail), and because the imaging data belongs to optical information, photoelectric conversion is required to be performed to convert the optical information into data to be evaluated which belongs to electrical information; based on the data to be evaluated, performing quality evaluation (such as evaluation of imaging face value, contrast, focal length, depth of field, field curvature and the like) on the zoom lens;
According to the embodiment of the invention, the quality evaluation is carried out on the zoom lens based on the data to be evaluated, so that a worker can be helped to rapidly evaluate the imaging quality of the lens in production, the imaging quality is improved, the reject ratio is reduced, and the enterprise cost is reduced by utilizing the evaluation result to carry out optical adjustment on the zoom lens.
The embodiment of the invention provides a zoom lens quality evaluation device, wherein the acquisition module 1 performs the following operations:
Providing a monochromatic light source;
And transmitting the light beams emitted by the monochromatic light sources to pass through the differentiation plate and the zoom lens in sequence, and imaging on a high-precision linear array CCD to obtain imaging data.
The working principle and the beneficial effects of the technical scheme are as follows:
As shown in fig. 2, a monochromatic light source is provided (for example, a light source machine is used); the light beam emitted by the monochromatic light source is transmitted through the dividing plate (the dividing plate has the function of superposing a cross wire or concentric ring pattern on an object to be imaged, the pattern can be used as a position reference and can be aligned with the object to be imaged, the cross dividing plate or concentric ring dividing plate can be used) and the zoom lens, imaging is carried out on the high-precision linear array CCD (Charge Coupled Device ) and the acquisition of imaging data is completed.
The embodiment of the invention provides a zoom lens quality evaluation device, wherein an evaluation module 2 performs the following operations:
Acquiring a preset evaluation target set, wherein the evaluation target set comprises: a plurality of first evaluation targets;
Acquiring attribute information of a zoom lens, and screening a second evaluation target to be evaluated from the first evaluation targets based on the attribute information;
Extracting target data corresponding to a second evaluation target from the data to be evaluated;
Determining an evaluation model corresponding to a second evaluation target based on a preset evaluation target-evaluation model library;
And inputting the target data into a corresponding evaluation model to obtain an evaluation result.
The working principle and the beneficial effects of the technical scheme are as follows:
The first evaluation target is specifically: for example, evaluating contrast, evaluating depth of field effect, and the like; the preset evaluation target-evaluation model library specifically comprises the following components: a database containing evaluation models corresponding to different evaluation targets, for example: the evaluation target is evaluation contrast, and the corresponding evaluation model is a model generated after a large number of records of manual evaluation contrast are learned by using a machine learning algorithm and is used for evaluating the contrast; the attribute information of the zoom lens specifically includes: for example, material provider information of the lens, production process of the lens, manufacturer, etc.;
After the second evaluation target is determined, target data (for example, contrast data) corresponding to the evaluation target in the data to be evaluated is determined, and the evaluation is performed by using a corresponding evaluation model.
The embodiment of the invention provides a zoom lens quality evaluation device, wherein an evaluation module 2 executes the following operations:
extracting a plurality of first attribute items in the attribute information;
acquiring the attribute type of the first attribute item, and determining an important value corresponding to the attribute type based on a preset attribute type-important value library;
if the importance value is greater than or equal to a preset importance value threshold value, the corresponding first attribute item is used as a second attribute item;
acquiring a preset negative event generation model, inputting a second attribute item into the negative event generation model, and acquiring at least one negative event and a first serious value corresponding to the negative event;
acquiring a preset capture strategy set, wherein the capture strategy set comprises: a plurality of capture strategies;
Determining at least one capturing object corresponding to the capturing strategy based on a preset capturing strategy-capturing object library;
attempting to capture at least one first lens quality event in the capture object based on the capture policy;
if the capturing is successful, acquiring a capturing process for capturing the first lens quality event;
carrying out flow splitting on the capturing flow to obtain a plurality of flows;
Extracting at least one first captured scene in the process;
Acquiring the credibility of a first capturing scene;
if the credibility is smaller than or equal to a preset credibility threshold, taking the corresponding first captured scene as a second captured scene;
attempting to acquire an association relationship between the second captured scene and the captured object;
If the acquisition fails, rejecting a corresponding first lens quality event;
If the acquisition is successful, analyzing the association relation to acquire a relation value;
If the relation value is smaller than or equal to a preset relation value threshold value, eliminating the corresponding first lens quality event;
When all the first lens quality events needing to be removed in the first lens quality events are removed, taking the remaining first lens quality events as second lens quality events;
Acquiring a preset receiving node set, wherein the receiving node set comprises: a plurality of receiving nodes;
On the basis of a preset query mode, the receiving node is queried at regular time, and at least one third shot quality event replied by the queried receiving node and at least one sender corresponding to the third shot quality event are acquired;
acquiring a sending record of a sender;
extracting a plurality of first record items in the transmission record;
Establishing a time axis, and setting the first record item on a corresponding time node on the time axis based on the generation time of the first record item;
Extracting features of the first record item to obtain a plurality of first features;
Acquiring a preset risk feature library, matching the first feature with a second feature in the risk feature library, taking the matched second feature as a third feature if the matching is met, and taking a corresponding first record item as a second record item;
Based on a preset feature-supplementing direction library, attempting to determine at least one supplementing direction corresponding to the third feature;
if the determination fails, randomly combining the first features obtained by extracting the features of the second record item to obtain a plurality of first combined features;
acquiring a preset malicious feature library, matching the first combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
If the determination is successful, selecting at least one first record item in a range preset in the supplementing direction of the second record item on the time axis, and taking the first record item as a third record item;
Extracting the characteristics of the third record item to obtain a plurality of fifth characteristics;
Randomly combining the first feature and the fifth feature obtained by extracting the features of the second record item to obtain a plurality of second combined features;
matching the second combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
When all the third lens quality events needing to be removed in the third lens quality events are removed, taking the rest third lens quality events as fourth lens quality events;
Acquiring a preset event determination model, determining whether a negative event is contained in a second lens quality event and a fourth lens quality event or not by the event determination model, and if so, outputting a first serious value corresponding to the contained negative event and taking the first serious value as a second serious value;
Summarizing the second serious value to obtain a sorting value;
Sequencing the first evaluation targets corresponding to the first attribute items according to the sizes of the corresponding sequencing values to obtain an evaluation target sequence;
and selecting first n first evaluation targets in the evaluation target sequence as second evaluation targets, and finishing screening.
The working principle and the beneficial effects of the technical scheme are as follows:
The preset attribute type-important value library specifically comprises the following steps: a database containing importance values corresponding to different attribute types, for example: the attribute type is a lens material provider, the corresponding importance value is 95, and the importance value represents the influence degree of the attribute type on the lens quality; the preset threshold value of the important value is specifically: for example, 75; the preset negative event generation model specifically comprises the following steps: a model generated after learning a number of manually made records of negative events using a machine learning algorithm, the model may generate negative events based on attribute terms, such as: the attribute item is a lens material manufacturer, the negative event is that the lens material produced by the manufacturer causes unqualified lens quality and the like, a first serious value of the negative event is output, and the larger the first serious value is, the more serious the corresponding negative event is; the capture strategy is specifically: for example, web crawling; the preset capturing strategy-capturing object library specifically comprises the following steps: data containing capture objects corresponding to different capture strategies, such as: the step strategy is that a webpage is crawled, and the capturing objects are a plurality of lens forum websites; the preset credibility threshold is specifically: for example 80; the preset relation value threshold value is specifically: for example, 90; the receiving node specifically comprises: a network node to which a peer can send data, i.e. complete data reception; the preset query mode specifically comprises the following steps: for example, if there is newly received data, if so, please reply; the preset risk feature library specifically comprises the following steps: a database containing a plurality of risk features, the risk features being features that may lead to malicious events, such as: the user wants to issue false data, suddenly changing identity information (risk feature); the preset malicious feature library specifically comprises the following steps: a database containing a plurality of malicious features, such as: changing identity information, and the released data is detected as false data; the preset range is specifically as follows: a range on the time axis corresponding to a certain length of time (e.g., 2 days); the preset event determination model specifically comprises the following steps: a model generated after the record of manually judging whether the event contains a certain event is learned by utilizing a machine learning algorithm, wherein the model can determine whether the event contains a certain event;
When screening the second evaluation target to be evaluated, it may be determined based on the attribute information of the zoom lens, for example: the manufacturers for producing the material of the zoom lens have the problem that the contrast of imaging of the lens is poor due to the material problem, and then the contrast evaluation is required; how to acquire problem events, however, needs to be resolved; the application extracts a first attribute item (such as a material provider) in attribute information, screens out a second attribute item which has influence on a quality result based on an important value, generates a negative event corresponding to the second attribute item based on a negative event generation model, and determines whether the negative event occurs or not; attempting to capture a shot quality event (for example, a quality evaluation article published by a user of a forum website) in a capture object, if capturing is successful, determining a plurality of processes in a capture process (for example, process 1, capturing a comment of "quality is not feasible" published by the user in a factory discussion area, process 2, capturing a homepage of the user to publish a specific article), wherein a first capture scene corresponding to the process is, for example: the process 1 corresponds to a web page of a discussion area of a manufacturer, and the process 2 corresponds to a web page of a user homepage; acquiring the credibility (webpage credibility) of the first captured scene, and taking the corresponding first captured scene as a second captured scene if the credibility is smaller than a credibility threshold; attempting to acquire the association relation between the second captured scene and the captured object (for example, the second captured scene belongs to a webpage of the captured object, the captured object guarantees the webpage to form a guarantee relation), if the acquisition fails, indicating that the second captured scene is not trusted (for example, from other forums), if the acquisition is successful, analyzing the relation (guarantee relation), acquiring a relation value, wherein the larger the relation value is, the larger the guarantee degree of the captured object guaranteeing the relation value is, and the reliability is high; some cooperative forums do not allow my to capture for privacy security, may set receiving nodes, and the counterpart may send to my; but when a third shot quality event is received, the corresponding sender needs to be verified; acquiring a sending record (record of data sent to a plurality of receiving nodes historically, etc.) of a sender, extracting a plurality of first record items, matching a first feature with a second feature, determining whether risk features possibly causing malicious events occur, if so, determining a supplementing direction, for example: the third feature is that the user changes the identity information suddenly, possibly uploading false data later, and the supplementing direction is later; if the supplementing direction cannot be determined, indicating that a malicious event exists in the second record item, respectively combining the first combination feature and the second combination feature based on the two conditions, and matching the first combination feature and the second combination feature with the fourth feature to determine whether the malicious event occurs; after eliminating the lens quality event to be eliminated, determining whether the rest lens quality event contains a corresponding negative event, wherein the larger the content is, the more the evaluation target corresponding to the first attribute item needs to be evaluated;
When the second evaluation target to be evaluated is selected from the first evaluation targets, the second evaluation targets can be determined based on the attribute information of the zoom lens, the setting is reasonable, the quality evaluation efficiency of the zoom lens is improved, and unnecessary workload is reduced; meanwhile, when acquiring the lens quality event, the unreliable lens quality event is removed, and the accuracy and safety of acquisition are ensured.
The embodiment of the invention provides a zoom lens quality evaluation device, which further comprises:
the expansion module is used for expanding the risk feature library;
The expansion module performs the following operations:
obtaining a preset extended node set, wherein the extended node set comprises: a plurality of first expansion nodes;
Acquiring the guarantee information of the first expansion node, wherein the guarantee information comprises the following components: a second extended node for guaranteeing the first extended node and a first guaranteeing value corresponding to the second extended node;
Determining a first expansion node in the second expansion nodes, taking the first expansion node as a third expansion node, and taking a first holding value corresponding to the third expansion node as a second holding value;
Taking a second expansion node except a third expansion node in the second expansion nodes as a fourth expansion node, and taking a first held value corresponding to the fourth expansion node as a third held value;
Acquiring a preset first calculation model, inputting the second guaranteed value and the third guaranteed value into the first calculation model, and acquiring a first score;
if the first score is greater than or equal to a preset first score threshold, the corresponding first expansion node is used as a fifth expansion node;
acquiring at least one first risk feature through a fifth expansion node;
Acquiring a preset forward verification model, inputting a first risk feature into the forward verification model, and acquiring at least one forward verification value;
Acquiring a preset reverse verification model, inputting the first risk characteristic into the reverse verification model, and acquiring at least one reverse verification value;
Acquiring a preset second calculation model, and inputting the forward verification value and the reverse verification value into the second calculation model to acquire a second score;
If the second score is greater than or equal to a preset second score threshold, the corresponding first risk feature is used as a second risk feature;
storing the second risk features into a risk feature library;
And after the second risk features which are required to be stored in the risk feature library are stored, completing expansion.
The working principle and the beneficial effects of the technical scheme are as follows:
The first expansion node specifically comprises: the network node corresponds to a collector of the risk characteristics, and the risk characteristics collected by the collector can be obtained through the node; the preset first scoring threshold value specifically comprises the following steps: for example, 85; the preset forward verification model specifically comprises the following steps: the machine learning algorithm is utilized to learn a plurality of manual forward verification records (based on a plurality of malicious event data, whether the risk features can positively cause the occurrence of malicious events or not) to generate a model, and the larger the forward verification value is, the larger the occurrence probability of the malicious events caused by the risk features is; the preset reverse verification model specifically comprises the following steps: the method comprises the steps of utilizing a machine learning algorithm to learn a model generated after a large number of manual reverse verification records (based on a large number of malicious event data, whether risk features exist before a malicious event happens or not is verified) and enabling the probability of the risk features to be larger before the malicious event happens to be larger as a reverse verification value is larger; the preset second scoring threshold value specifically comprises the following steps: for example, 90;
when the first expansion node is collected Fang Duiying, other second expansion nodes are required to guarantee the first expansion node (guarantee is carried out, one side generates bad records, the other side also generates influences, for example, the guarantee value is reduced); calculating a first score based on the second guaranteed value and the third guaranteed value, wherein the higher the first score is, the higher the credibility of the corresponding first expansion node is; when the first risk feature is acquired, whether the risk feature is real or not needs to be verified in a forward and reverse direction, and a second risk feature meeting the condition is screened out to expand a risk feature library; the efficiency of the system for finding risk features can be further improved;
The preset first calculation model specifically comprises the following steps: a model with a built-in calculation formula, the calculation formula of which is as follows:
Wherein σ is the first score, A i is the i-th second held value, l is the total number of the second held values, B i is the i-th third held value, d is the total number of the third held values, ε 1 and ε 2 are preset weight values, Ρ is an intermediate variable;
In the formula, the second guarantee value and the third guarantee value are positively correlated with the first score, and because the third expansion node approved by the user corresponding to the second guarantee value guarantees the first expansion node to acquire the higher the guarantee approval degree, the l-d is positively correlated with the first score;
the preset second calculation model specifically comprises the following steps: a model with a built-in calculation formula, the calculation formula of which is as follows:
Wherein γ is the second score, e is a natural constant, α t is the t-th forward verification value, n 1 is the total number of the forward verification values, β t is the t-th reverse verification value, n 2 is the total number of the reverse verification values, ε is the sum of the first number of the forward verification values less than or equal to a preset first threshold value and the second number of the reverse verification values less than or equal to a preset second threshold value in the forward verification values;
In the formula, the positive verification value and the negative verification value are positively correlated with the second score, the first number and the second number are negatively correlated with the second score, and the sum of the first number and the second number is negatively correlated with the second score.
Through the formula, the first score and the second score are calculated rapidly, the expansion nodes and the risk features are convenient to screen, and the working efficiency of the system is improved to a great extent.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (4)
1. A zoom lens quality evaluation method, characterized by comprising:
Acquiring imaging data of a zoom lens, and performing photoelectric conversion on the imaging data to obtain data to be evaluated;
Based on the data to be evaluated, performing quality evaluation on the zoom lens;
based on the data to be evaluated, performing quality evaluation on the zoom lens, including:
acquiring a preset evaluation target set, wherein the evaluation target set comprises: a plurality of first evaluation targets;
Acquiring attribute information of the zoom lens, and screening a second evaluation target which needs to be evaluated from the first evaluation targets based on the attribute information;
extracting target data corresponding to the second evaluation target from the data to be evaluated;
determining an evaluation model corresponding to the second evaluation target based on a preset evaluation target-evaluation model library;
Inputting the target data into a corresponding evaluation model to obtain an evaluation result;
and screening a second evaluation target to be evaluated from the first evaluation targets based on the attribute information, wherein the second evaluation target comprises:
Extracting a plurality of first attribute items in the attribute information;
Acquiring the attribute type of the first attribute item, and determining an important value corresponding to the attribute type based on a preset attribute type-important value library;
If the importance value is greater than or equal to a preset importance value threshold, the corresponding first attribute item is used as a second attribute item;
acquiring a preset negative event generation model, inputting the second attribute item into the negative event generation model, and acquiring at least one negative event and a first serious value corresponding to the negative event;
acquiring a preset capture strategy set, wherein the capture strategy set comprises: a plurality of capture strategies;
determining at least one capturing object corresponding to a capturing strategy based on a preset capturing strategy-capturing object library;
attempting to capture at least one first lens quality event in the capture object based on the capture policy;
If the capturing is successful, acquiring a capturing process for capturing the first lens quality event;
performing flow splitting on the capturing flow to obtain a plurality of flows;
Extracting at least one first captured scene in the flow;
Acquiring the credibility of the first captured scene;
if the credibility is smaller than or equal to a preset credibility threshold, taking the corresponding first captured scene as a second captured scene;
attempting to acquire the association relation between the second captured scene and the captured object;
If the acquisition fails, rejecting the corresponding first lens quality event;
if the acquisition is successful, analyzing the association relation to acquire a relation value;
If the relation value is smaller than or equal to a preset relation value threshold value, eliminating the corresponding first lens quality event;
When all the first lens quality events needing to be removed in the first lens quality events are removed, taking the rest first lens quality events as second lens quality events;
Acquiring a preset receiving node set, wherein the receiving node set comprises: a plurality of receiving nodes;
Based on a preset query mode, the receiving node is queried at regular time, and at least one third shot quality event replied by the queried receiving node and at least one sender corresponding to the third shot quality event are obtained;
Acquiring a sending record of the sender;
extracting a plurality of first record items in the sending record;
establishing a time axis, and setting a first record item on a corresponding time node on the time axis based on the generation time of the first record item;
extracting the characteristics of the first record item to obtain a plurality of first characteristics;
acquiring a preset risk feature library, matching the first feature with a second feature in the risk feature library, taking the second feature matched with the first feature as a third feature if the first feature is matched with the second feature, and taking the corresponding first record item as a second record item;
based on a preset feature-supplementing direction library, attempting to determine at least one supplementing direction corresponding to the third feature;
if the first record item is determined to be failed, randomly combining the first features obtained by extracting the features of the second record item to obtain a plurality of first combined features;
acquiring a preset malicious feature library, matching the first combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
If the determination is successful, selecting at least one first record item in a range preset in the supplementing direction of the second record item on the time axis, and taking the first record item as a third record item;
Extracting the characteristics of the third record item to obtain a plurality of fifth characteristics;
randomly combining the first feature and the fifth feature obtained by extracting the features of the second record item to obtain a plurality of second combined features;
Matching the second combined feature with a fourth feature in the malicious feature library, and eliminating the corresponding third lens quality event if the matching is met;
when all the third lens quality events needing to be removed in the third lens quality events are removed, taking the rest third lens quality events as fourth lens quality events;
acquiring a preset event determination model, determining whether the second lens quality event and the fourth lens quality event contain the negative event or not by using the event determination model, and if so, outputting the first serious value corresponding to the contained negative event and taking the first serious value as a second serious value;
Summarizing the second serious value to obtain a sorting value;
sorting the first evaluation targets corresponding to the first attribute items according to the sorting values from large to small to obtain an evaluation target sequence;
Selecting the first n first evaluation targets in the evaluation target sequence as second evaluation targets, and finishing screening;
The method further comprises the steps of:
Expanding the risk feature library;
wherein, expand the said risk feature library, including:
acquiring a preset extended node set, wherein the extended node set comprises: a plurality of first expansion nodes;
acquiring the guarantee information of the first expansion node, wherein the guarantee information comprises: a second extended node that vouches for the first extended node and a first vouch-for value corresponding to the second extended node;
Determining the first expansion node in the second expansion nodes and taking the first expansion node as a third expansion node, and simultaneously taking the first holding value corresponding to the third expansion node as a second holding value;
Taking the second expansion nodes except the third expansion node in the second expansion nodes as fourth expansion nodes, and taking the first held value corresponding to the fourth expansion nodes as a third held value;
acquiring a preset first calculation model, inputting the second guarantee value and the third guarantee value into the first calculation model, and acquiring a first score;
If the first score is greater than or equal to a preset first score threshold, the corresponding first expansion node is used as a fifth expansion node;
Acquiring at least one first risk feature through the fifth expansion node;
Acquiring a preset forward verification model, inputting the first risk characteristic into the forward verification model, and acquiring at least one forward verification value;
Acquiring a preset reverse verification model, inputting the first risk characteristic into the reverse verification model, and acquiring at least one reverse verification value;
acquiring a preset second calculation model, and inputting the forward verification value and the reverse verification value into the second calculation model to acquire a second score;
If the second score is greater than or equal to a preset second score threshold, the corresponding first risk feature is used as a second risk feature;
storing the second risk feature into the risk feature library;
And after the second risk features which are required to be stored in the risk feature library are stored, completing expansion.
2. The method of evaluating quality of a zoom lens according to claim 1, wherein acquiring imaging data of the zoom lens comprises:
Providing a monochromatic light source;
And transmitting the light beams emitted by the monochromatic light sources to pass through the differentiation plate and the zoom lens in sequence, and imaging on a high-precision linear array CCD to obtain imaging data.
3. A zoom lens quality evaluation apparatus, comprising:
the acquisition module is used for acquiring imaging data of the zoom lens, and performing photoelectric conversion on the imaging data to acquire data to be evaluated;
The evaluation module is used for evaluating the quality of the zoom lens based on the data to be evaluated;
the evaluation module performs the following operations:
acquiring a preset evaluation target set, wherein the evaluation target set comprises: a plurality of first evaluation targets;
Acquiring attribute information of the zoom lens, and screening a second evaluation target which needs to be evaluated from the first evaluation targets based on the attribute information;
extracting target data corresponding to the second evaluation target from the data to be evaluated;
determining an evaluation model corresponding to the second evaluation target based on a preset evaluation target-evaluation model library;
Inputting the target data into a corresponding evaluation model to obtain an evaluation result;
the evaluation module performs the following operations:
Extracting a plurality of first attribute items in the attribute information;
Acquiring the attribute type of the first attribute item, and determining an important value corresponding to the attribute type based on a preset attribute type-important value library;
If the importance value is greater than or equal to a preset importance value threshold, the corresponding first attribute item is used as a second attribute item;
acquiring a preset negative event generation model, inputting the second attribute item into the negative event generation model, and acquiring at least one negative event and a first serious value corresponding to the negative event;
acquiring a preset capture strategy set, wherein the capture strategy set comprises: a plurality of capture strategies;
determining at least one capturing object corresponding to a capturing strategy based on a preset capturing strategy-capturing object library;
attempting to capture at least one first lens quality event in the capture object based on the capture policy;
If the capturing is successful, acquiring a capturing process for capturing the first lens quality event;
performing flow splitting on the capturing flow to obtain a plurality of flows;
Extracting at least one first captured scene in the flow;
Acquiring the credibility of the first captured scene;
if the credibility is smaller than or equal to a preset credibility threshold, taking the corresponding first captured scene as a second captured scene;
attempting to acquire the association relation between the second captured scene and the captured object;
If the acquisition fails, rejecting the corresponding first lens quality event;
if the acquisition is successful, analyzing the association relation to acquire a relation value;
If the relation value is smaller than or equal to a preset relation value threshold value, eliminating the corresponding first lens quality event;
When all the first lens quality events needing to be removed in the first lens quality events are removed, taking the rest first lens quality events as second lens quality events;
Acquiring a preset receiving node set, wherein the receiving node set comprises: a plurality of receiving nodes;
Based on a preset query mode, the receiving node is queried at regular time, and at least one third shot quality event replied by the queried receiving node and at least one sender corresponding to the third shot quality event are obtained;
Acquiring a sending record of the sender;
extracting a plurality of first record items in the sending record;
establishing a time axis, and setting a first record item on a corresponding time node on the time axis based on the generation time of the first record item;
extracting the characteristics of the first record item to obtain a plurality of first characteristics;
acquiring a preset risk feature library, matching the first feature with a second feature in the risk feature library, taking the second feature matched with the first feature as a third feature if the first feature is matched with the second feature, and taking the corresponding first record item as a second record item;
based on a preset feature-supplementing direction library, attempting to determine at least one supplementing direction corresponding to the third feature;
if the first record item is determined to be failed, randomly combining the first features obtained by extracting the features of the second record item to obtain a plurality of first combined features;
acquiring a preset malicious feature library, matching the first combined feature with a fourth feature in the malicious feature library, and eliminating a corresponding third lens quality event if the matching is met;
If the determination is successful, selecting at least one first record item in a range preset in the supplementing direction of the second record item on the time axis, and taking the first record item as a third record item;
Extracting the characteristics of the third record item to obtain a plurality of fifth characteristics;
randomly combining the first feature and the fifth feature obtained by extracting the features of the second record item to obtain a plurality of second combined features;
Matching the second combined feature with a fourth feature in the malicious feature library, and eliminating the corresponding third lens quality event if the matching is met;
when all the third lens quality events needing to be removed in the third lens quality events are removed, taking the rest third lens quality events as fourth lens quality events;
acquiring a preset event determination model, determining whether the second lens quality event and the fourth lens quality event contain the negative event or not by using the event determination model, and if so, outputting the first serious value corresponding to the contained negative event and taking the first serious value as a second serious value;
Summarizing the second serious value to obtain a sorting value;
sorting the first evaluation targets corresponding to the first attribute items according to the sorting values from large to small to obtain an evaluation target sequence;
Selecting the first n first evaluation targets in the evaluation target sequence as second evaluation targets, and finishing screening;
The apparatus further comprises:
the expansion module is used for expanding the risk feature library;
the expansion module performs the following operations:
acquiring a preset extended node set, wherein the extended node set comprises: a plurality of first expansion nodes;
acquiring the guarantee information of the first expansion node, wherein the guarantee information comprises: a second extended node that vouches for the first extended node and a first vouch-for value corresponding to the second extended node;
Determining the first expansion node in the second expansion nodes and taking the first expansion node as a third expansion node, and simultaneously taking the first holding value corresponding to the third expansion node as a second holding value;
Taking the second expansion nodes except the third expansion node in the second expansion nodes as fourth expansion nodes, and taking the first held value corresponding to the fourth expansion nodes as a third held value;
acquiring a preset first calculation model, inputting the second guarantee value and the third guarantee value into the first calculation model, and acquiring a first score;
If the first score is greater than or equal to a preset first score threshold, the corresponding first expansion node is used as a fifth expansion node;
Acquiring at least one first risk feature through the fifth expansion node;
Acquiring a preset forward verification model, inputting the first risk characteristic into the forward verification model, and acquiring at least one forward verification value;
Acquiring a preset reverse verification model, inputting the first risk characteristic into the reverse verification model, and acquiring at least one reverse verification value;
acquiring a preset second calculation model, and inputting the forward verification value and the reverse verification value into the second calculation model to acquire a second score;
If the second score is greater than or equal to a preset second score threshold, the corresponding first risk feature is used as a second risk feature;
storing the second risk feature into the risk feature library;
And after the second risk features which are required to be stored in the risk feature library are stored, completing expansion.
4. A zoom lens quality evaluation apparatus according to claim 3, wherein the acquisition module performs the operations of:
Providing a monochromatic light source;
And transmitting the light beams emitted by the monochromatic light sources to pass through the differentiation plate and the zoom lens in sequence, and imaging on a high-precision linear array CCD to obtain imaging data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111170279.4A CN113923444B (en) | 2021-10-08 | 2021-10-08 | Zoom lens quality evaluation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111170279.4A CN113923444B (en) | 2021-10-08 | 2021-10-08 | Zoom lens quality evaluation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113923444A CN113923444A (en) | 2022-01-11 |
| CN113923444B true CN113923444B (en) | 2024-04-30 |
Family
ID=79237984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111170279.4A Active CN113923444B (en) | 2021-10-08 | 2021-10-08 | Zoom lens quality evaluation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113923444B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102735429A (en) * | 2012-06-13 | 2012-10-17 | 中国科学院长春光学精密机械与物理研究所 | Equipment for CCD (Charge Coupled Device) modulation transfer function test and testing method of equipment |
| CN108989789A (en) * | 2018-07-11 | 2018-12-11 | 深圳职业技术学院 | A kind of camera imaging quality real-time estimating method |
| CN109521547A (en) * | 2018-12-21 | 2019-03-26 | 广州医软智能科技有限公司 | A kind of automatic focusing method and system of variable step |
| CN109883656A (en) * | 2019-03-26 | 2019-06-14 | 北京全欧光学检测仪器有限公司 | The non-detection device and method for improving imaging lens |
| CN111062378A (en) * | 2019-12-23 | 2020-04-24 | 重庆紫光华山智安科技有限公司 | Image processing method, model training method, target detection method and related device |
| CN111432125A (en) * | 2020-03-31 | 2020-07-17 | 合肥英睿系统技术有限公司 | Focusing method and device, electronic equipment and storage medium |
| CN112529871A (en) * | 2020-12-11 | 2021-03-19 | 杭州海康威视系统技术有限公司 | Method and device for evaluating image and computer storage medium |
| CN112911281A (en) * | 2021-02-09 | 2021-06-04 | 北京三快在线科技有限公司 | Video quality evaluation method and device |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8089516B2 (en) * | 2006-06-20 | 2012-01-03 | Hewlett-Packard Development Company, L.P. | Event management for camera systems |
| JP6624878B2 (en) * | 2015-10-15 | 2019-12-25 | キヤノン株式会社 | Image processing apparatus, image processing method, and program |
-
2021
- 2021-10-08 CN CN202111170279.4A patent/CN113923444B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102735429A (en) * | 2012-06-13 | 2012-10-17 | 中国科学院长春光学精密机械与物理研究所 | Equipment for CCD (Charge Coupled Device) modulation transfer function test and testing method of equipment |
| CN108989789A (en) * | 2018-07-11 | 2018-12-11 | 深圳职业技术学院 | A kind of camera imaging quality real-time estimating method |
| CN109521547A (en) * | 2018-12-21 | 2019-03-26 | 广州医软智能科技有限公司 | A kind of automatic focusing method and system of variable step |
| CN109883656A (en) * | 2019-03-26 | 2019-06-14 | 北京全欧光学检测仪器有限公司 | The non-detection device and method for improving imaging lens |
| CN111062378A (en) * | 2019-12-23 | 2020-04-24 | 重庆紫光华山智安科技有限公司 | Image processing method, model training method, target detection method and related device |
| CN111432125A (en) * | 2020-03-31 | 2020-07-17 | 合肥英睿系统技术有限公司 | Focusing method and device, electronic equipment and storage medium |
| CN112529871A (en) * | 2020-12-11 | 2021-03-19 | 杭州海康威视系统技术有限公司 | Method and device for evaluating image and computer storage medium |
| CN112911281A (en) * | 2021-02-09 | 2021-06-04 | 北京三快在线科技有限公司 | Video quality evaluation method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113923444A (en) | 2022-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tsintotas et al. | Assigning visual words to places for loop closure detection | |
| Raguram et al. | USAC: A universal framework for random sample consensus | |
| David et al. | Softposit: Simultaneous pose and correspondence determination | |
| CN110473179B (en) | Method, system and equipment for detecting surface defects of thin film based on deep learning | |
| Feng et al. | Cityflow-nl: Tracking and retrieval of vehicles at city scale by natural language descriptions | |
| CN112381075A (en) | Method and system for carrying out face recognition under specific scene of machine room | |
| TWI709188B (en) | Fusion-based classifier, classification method, and classification system | |
| EP4139892A2 (en) | Method and apparatus for camera calibration | |
| CN112116582A (en) | Cigarette detection and identification method under stock or display scene | |
| CN113923444B (en) | Zoom lens quality evaluation method and device | |
| CN115188066A (en) | Moving target detection system and method based on cooperative attention and multi-scale fusion | |
| CN115379308B (en) | Internet of things equipment data acquisition system based on satellite remote communication | |
| Bammey | Analysis and experimentation on the ManTraNet image forgery detector | |
| JP2020013594A (en) | Information processing method, program, and information processing device | |
| Xue et al. | Detection of Various Types of Metal Surface Defects Based on Image Processing. | |
| CN110807453A (en) | OCR-based product character online detection method, device and system | |
| Xu et al. | Stereo camera trap for wildlife in situ observations and measurements | |
| Tan et al. | An application of an improved FCOS algorithm in detection and recognition of industrial instruments | |
| CN115587297A (en) | Method, apparatus, device and medium for constructing image recognition model and image recognition | |
| CN115908886A (en) | Image classification method, image processing apparatus, and storage device | |
| CN111813987B (en) | Portrait comparison method based on police big data | |
| KR20230008810A (en) | Create a panorama with your mobile camera | |
| CN113743443A (en) | Image evidence classification and identification method and device | |
| CN113111888A (en) | Picture distinguishing method and device | |
| CN110415128A (en) | Policy information management method, device, equipment and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |