CN118781062B

CN118781062B - An image detection system based on automation

Info

Publication number: CN118781062B
Application number: CN202410829328.8A
Authority: CN
Inventors: 朱静; 邵俊豪; 倪圣健; 张新年; 陈萌萌; 易星星; 杨德佩
Original assignee: Zhongtong Weiyi Technology Service Co ltd
Current assignee: China Telecom Service Testing and Certification Co., Ltd.
Priority date: 2024-06-25
Filing date: 2024-06-25
Publication date: 2025-07-01
Anticipated expiration: 2044-06-25
Also published as: CN118781062A

Abstract

The application discloses an automatic-based imager detection system, relates to the technical field of imagers, and solves the technical problems that the adjustment precision and efficiency of a light source module in the prior art are insufficient, and the imaging quality and detection efficiency of target object detection are affected; partitioning a target to be detected based on characteristic data of a target object and light source characteristics to obtain a plurality of detection areas; the application adjusts the light source according to the characteristic data of each target object to improve the quality of the collected image data, acquires the illumination characteristic model which can represent the mapping relation among the material characteristic, the color characteristic and the corresponding optimal illumination parameter, and acquires the illumination parameter corresponding to each detection area through the illumination characteristic model, and acquires the illumination parameter of each detection area through the artificial intelligent model to improve the accuracy and the image quality of the light source adjustment.

Description

Automatic-based imager detection system

Technical Field

The application belongs to the field of imagers, relates to an automatic detection technology of an imager, and particularly relates to an automatic-based imager detection system.

Background

The full-automatic image measuring instrument is an artificial intelligent modern optical non-contact measuring instrument developed on the basis of a digital image measuring instrument (also called CNC image instrument). The full-automatic image measuring instrument has excellent motion precision and motion control performance of a digital instrument, integrates design flexibility of machine software, and belongs to current leading edge optical size detection equipment.

When the existing imager detection system detects and analyzes a target object, a plurality of sets of hardware equipment of the imager are required to cooperate to acquire image data of the target object, and then various parameters of the target object are obtained according to the image data analysis. The light source module is crucial in the image data acquisition process, the imaging quality is required to be ensured by adaptively adjusting the illumination parameters, but under the general condition, the illumination parameters of the light source module are complicated to adjust, the imaging quality of different areas of the same target object cannot be ensured, and the detection efficiency of continuous different target objects can be influenced.

The application provides an automatic-based imager detection system for solving the technical problems.

Disclosure of Invention

The application aims to at least solve one of the technical problems in the prior art, and therefore, the application provides an automatic-based imager detection system which is used for solving the technical problems that the adjustment precision and efficiency of a light source module in the prior art are insufficient, and the imaging quality and detection efficiency of target object detection are affected.

In order to achieve the above object, a first aspect of the present application provides an automated imager detection system, comprising a light source adjusting module and a data acquisition module connected with the light source adjusting module;

The data acquisition module is used for extracting characteristic data of each target object in the target group to be detected by combining the original shape, correlating the characteristic data with the detection sequence of the target objects to generate a characteristic sequence, wherein the characteristic data comprises the original shape, and material characteristics and color characteristics corresponding to each position in the original shape;

The light source adjusting module is used for determining a target to be detected according to the characteristic sequence, extracting characteristic data of the target to be detected, partitioning the target to be detected based on the characteristic data and the light source characteristics to obtain a plurality of detection areas, and

The method comprises the steps of constructing an illumination characteristic model, acquiring corresponding illumination parameters by combining characteristic data of a plurality of detection areas, adjusting light sources according to the illumination parameters of the plurality of detection areas, and further acquiring image data of an object to be detected, wherein the illumination characteristic model is constructed based on an artificial intelligent model.

When detecting the target object based on the imager, various parameters of the target object, such as size, shape, surface defects and the like, are analyzed mainly by collecting image data of the target object and combining with a built-in analysis algorithm. Therefore, the quality of the image data is of paramount importance. However, during image acquisition, the configuration and adjustment of the light sources can have a significant impact on image quality. The target objects detected by the imager are often various, and the prior art is difficult to automatically adjust the light sources aiming at different target objects, so that high-quality image data are difficult to acquire.

The application firstly determines basic characteristics of each target object in a group to be detected, and associates the basic characteristics with the original shape of the target object to obtain characteristic data. And then when the target to be detected is detected, dividing the target to be detected into a plurality of detection areas according to the characteristic data and the light source characteristics. And analyzing the most appropriate illumination parameters of each detection area through the illumination characteristic model, and carrying out partition adjustment on the light source through the illumination parameters in the detection process so as to ensure the quality of the acquired image data.

Preferably, extracting feature data of each target object in the target group to be detected by combining the original shape includes:

obtaining basic characteristics and original shapes of target objects in a target group to be detected through a database, wherein the basic characteristics comprise material characteristics and color characteristics;

Dividing the original shape of the target object into a plurality of first characteristic groups according to the material characteristics, and dividing the first characteristic groups into a plurality of second characteristic groups according to the color characteristics;

and associating the second characteristic groups with the corresponding positions of the second characteristic groups in the original shape to obtain characteristic data.

In the working process of the imager, the image quality can be influenced by the material, the material and the like of the target object. To improve the image quality by adjusting the light source by division, it is first necessary to divide the target object, and the division is based on factors affecting the image quality, so that the characteristics of each position of the target object are clarified.

The application firstly extracts basic characteristics of each target object in the target group to be detected, namely material characteristics and color characteristics of the target object. Firstly dividing the original shape of the target object into a plurality of first feature groups according to different materials, then dividing the first feature groups into a plurality of second feature groups according to the difference of colors, dividing each second feature group into a plurality of second feature groups according to the materials and the colors, and associating each second feature group with the position of each second feature group in the original shape of the target object to obtain feature data.

After the characteristic data of the target to be detected are obtained, the material and color characteristics at each position of the target can be quickly identified, and a data foundation is laid for dividing the target object into a plurality of detection areas.

Preferably, partitioning the object to be detected based on the feature data and the light source features includes:

extracting light source characteristics of the light source module, and determining partition quantity and calibration line rings according to the light source characteristics, wherein the calibration line rings are determined according to the shape of a target object or the light source characteristics;

sequentially marking target objects in the characteristic sequence as targets to be detected, and calculating the comprehensive similarity of the characteristic data of the targets to be detected at adjacent positions on the calibration line ring, wherein the comprehensive similarity comprises material similarity and color similarity;

A number of detection zones is determined based on the integrated similarity and the number of zones.

When dividing the target object, firstly, whether the light source module can be controlled in a dividing way or not and can be independently controlled in a plurality of dividing ways is considered, and further, the reference of which is used for dividing the target object is considered.

The application acquires the partition number after determining that the light source module can be controlled in a partition mode. The calibration line ring is determined by the height between the light source module and the workbench and the included angle between the light and the workbench, and the size of the target object is considered when the calibration line ring is determined. The calibration wire loop can be set to be round if the target object is round, and the calibration wire loop can be set to be square if the target object is square. The light source characteristics comprise the shape of the light source and how many control areas the light source can be divided into, and when the shape of the light source is round, the calibration wire ring can also be set to be round.

And calculating the comprehensive similarity of the characteristic data of the object to be detected at the adjacent positions on the calibrated wire loop. And segmenting the calibration wire ring based on the comprehensive similarity and the partition number, so as to realize the partition of the target object. The application can ensure that the materials and the colors in each detection area are similar, is beneficial to adjusting the light source responsible for the detection area, and can also ensure that the quality of the image data obtained after adjustment is better.

Preferably, calculating the comprehensive similarity of the feature data of the object to be detected at the adjacent position on the calibrated wire loop includes:

dividing the calibration wire ring into a plurality of wire ring segments according to a set step length;

The method comprises the steps of marking the material similarity and the color similarity of adjacent wire ring segments as CXD and YXD, calculating the comprehensive similarity ZXD through a formula ZXD=α1×CXD+α2×YXD, wherein α1 and α2 are weight coefficients, α1 is larger than or equal to α2, and the material similarity is used for evaluating whether the reflection characteristics of the materials are similar or not.

According to the method, the calibrated wire ring is segmented to obtain a plurality of wire ring segments, and then the comprehensive similarity of the adjacent wire ring segments is calculated. The calculation of the comprehensive similarity is mainly based on the characteristic data of the corresponding area of each wire loop segment, and is mainly based on materials, colors and the like to judge whether the light source parameters required by the adjacent wire loop segments are similar or not. According to the application, whether the adjacent wire ring segments are similar is calculated through the characteristic data, and the method is used for judging whether the requirements of the adjacent wire ring segments on the light source parameters are similar, and if so, the adjacent wire ring segments can be combined.

Preferably, determining a number of detection zones based on the integrated similarity and the number of zones includes:

Calculating and determining the partition area according to the partition number;

When the comprehensive similarity of the adjacent wire ring segments is larger than a similarity threshold value, judging whether the combined area of the adjacent wire ring segments exceeds the partition area, if so, not combining the adjacent wire ring segments, otherwise, combining the adjacent wire ring segments;

and (5) adjusting the combination result to obtain a plurality of detection areas.

After the number of partitions is determined, it is also necessary to determine the partition area corresponding to each partition according to the height of the workbench and the light source module, and divide the target object into a plurality of detection areas by the partition area.

The method comprises the steps of judging whether the comprehensive similarity is larger than a similarity threshold value, merging target object areas corresponding to adjacent wire ring segments if the comprehensive similarity is larger than the similarity threshold value, judging whether the target object areas are larger than the partition area, if so, not merging, and if not, merging the adjacent wire ring segments to form a new wire ring segment. And comparing the new wire loop segment with the next adjacent wire loop segment to judge whether merging is needed.

Preferably, the obtaining the corresponding illumination parameters by combining the feature data of the plurality of detection areas includes:

sequentially extracting material characteristics and color characteristics from the characteristic data of the detection area;

The method comprises the steps of integrating material characteristics and color characteristics into model input data of an illumination characteristic model, inputting the model input data into the illumination characteristic model to obtain illumination parameters of a detection area, wherein the illumination parameters comprise illumination intensity and color temperature.

The material characteristics, the color characteristics and the illumination parameters are difficult to express by a very clear mapping relation, so that the application is realized by adopting an artificial intelligent model with strong nonlinear fitting capability. And extracting material characteristics and color characteristics from the characteristic data of each detection area, integrating the material characteristics and the color characteristics, and inputting the integrated material characteristics and the color characteristics into a constructed illumination adjustment model to obtain illumination parameters corresponding to the detection areas.

Preferably, constructing the illumination feature model includes:

the method comprises the steps of obtaining standard training data, wherein the standard training data comprises standard input data and standard output data, the standard input data is integrated based on material characteristics and color characteristics, and the standard input data is corresponding illumination parameters;

training an artificial intelligent model through standard training data, and marking the trained artificial intelligent model as an illumination characteristic model, wherein the standard training data is obtained through a large number of scene simulations.

When the artificial intelligent model is trained, simulation experiments are required to be carried out on different material characteristics and color characteristics of the detection area, so that the proper illumination parameters of the artificial intelligent model are obtained. After the multiple sets of data are obtained through simulation, the multiple sets of data sets can be preprocessed to obtain standard training data. Preprocessing comprises abnormal data rejection, data expansion and the like. And training the artificial intelligent model through standard training data to obtain the illumination characteristic model.

Preferably, the adjusting the light source according to the illumination parameters of the detection areas includes:

extracting a plurality of detection areas and corresponding illumination parameters of a target to be detected;

when the target to be detected enters the image acquisition area, the light source is adjusted through the illumination parameters of each detection area, and the image data of the target to be detected is acquired after the adjustment is completed.

After a plurality of detection areas and corresponding illumination parameters of each target to be detected are obtained, once the target to be detected enters an image acquisition area (arranged on a workbench), the light source is subjected to partition adjustment based on the illumination parameters, so that illumination of each detection area of the target to be detected can be ensured to acquire high-quality image data.

Compared with the prior art, the application has the beneficial effects that:

1. the application sequentially extracts each target object from the target group to be detected, partitions the target to be detected based on the characteristic data and the light source characteristics of the target object to obtain a plurality of detection areas, determines the corresponding optimal illumination parameters according to the material characteristics and the color characteristics of each detection area, and further completes the light source adjustment in the detection process.

2. According to the application, the artificial intelligent model is trained through a large amount of standard training data to obtain the illumination characteristic model capable of representing the mapping relation among the material characteristics, the color characteristics and the corresponding optimal illumination parameters, and the illumination parameters corresponding to all detection areas are obtained through the illumination characteristic model.

Drawings

In order to more clearly illustrate the embodiments of the application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of a system according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a gear detection area I according to an embodiment of the present application;

FIG. 3 is a second schematic diagram of a gear detection area according to an embodiment of the present application;

FIG. 4 is a schematic diagram of a clock detection region according to an embodiment of the application;

FIG. 5 is a flow chart of a method according to an embodiment of the application.

Detailed Description

The technical solutions of the present application will be clearly and completely described in connection with the embodiments, and it is obvious that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

Referring to fig. 1-5, an embodiment of a first aspect of the present application provides an automated-based imager detection system, including a light source adjustment module and a data acquisition module connected thereto;

The imaging instrument mainly comprises the following hardware:

The image module adopts a high-resolution color CCD of an original imported chip of Sony Japan, has excellent imaging, definition and reality and full color. The optical lens adopts an imported high-quality optical component, and is bright and clear after being subjected to multilayer optical coating, oil stain resistance, corrosion resistance, tiny optical attenuation and long-term use.

The light source module is designed into an inlet LED annular cold light source six-ring eight-zone design, is subjected to stepless adjustment, and has fine and soft light rays and bright and clear surface light imaging. The imaging is clear, the corners are clear, and the measuring is more accurate particularly for shaft-like and arc-like workpieces. The light source adjusting module is used for mainly adjusting and controlling the light source of the light source module.

And the transmission module is a CNC full-closed-loop controller (four-axis), a high-precision full-closed-loop servo motor, an integrated C3-level configured double-angle contact bearing and a grinding C3-level ball screw. Full closed loop control, high-speed response, accurate positioning and stable operation.

The lubricating module comprises an oil way distributor, a screw rod nut, a guide rail slide block oil injector, a high-pressure oil pipe and special oil for a Japanese THK guide rail screw rod. The oil way design of the system is standardized, and the high-quality grease ensures that the instrument can keep lasting precision and running stability under the conditions of long-term use and high-speed running.

The shock absorption module is used for enabling the instrument to obtain higher precision, speed and stability under the use environment of vibration or under the high-speed running state.

The glass table is made of German SCHOTT import raw sheet glass by a micro-float process, and has extremely high flatness and light transmittance after being precisely ground.

The workbench is a three-layer workbench design, and the high-strength aviation aluminum alloy material has super-strong hardness and stability in and on the surface of the material through a special treatment process. The long natural aging process makes the workbench eliminate the change of internal stress and has very stable physical properties.

The base stand column, the base and the stand column adopt high-precision natural granite, the 00-level grinding treatment is adopted, the physical characteristics are stable, no internal stress changes exist, the triaxial has the same temperature characteristics and expansion coefficient, and the durability, the reliability and the stability of the instrument precision are ensured.

Before the light source adjustment is performed, it is necessary to determine which target objects are detected and which target objects have characteristics that affect the light source adjustment.

The application firstly extracts a target group to be detected from a database, wherein the target group to be detected comprises a plurality of target objects which need to be detected by an imager, and basic characteristics and original shapes of all the target objects. The original shape is the outline shape of the target object, such as the original shape of a gear is a round cake shape, and the basic characteristics are some characteristics which influence the adjustment of a light source, such as the reflection of light by materials can influence the image quality, and the color of the object can also influence the image quality. Therefore, the target objects in the group to be detected are detection objects, and the basic characteristics of each target object are key for influencing illumination adjustment.

After the definition of the above, the original shape of the target object is firstly divided according to the characteristics of the materials, namely, the areas with the same materials are marked in the original shape, each mark is used as a first characteristic group, and the first characteristic groups are divided according to the color characteristics to obtain second characteristic groups. The corresponding position of each second feature group is different from the corresponding position of other second feature groups by at least one feature (material feature or color feature).

And associating the divided second feature group with the position of the second feature group in the original shape, namely associating one second feature group with each position in the original shape. In principle, the illumination parameters of the two corresponding positions of each feature set should be adjusted independently, but the application also needs to combine the positions in the original shape of the target object to obtain a plurality of detection areas in consideration of the number of partitions of the light source module and the influence of light source interference.

It is noted that the image data acquired by the imager is used for analyzing the size, the defect and the like of the target object by default, and the influence of the reflection intensity of the analysis material is larger, so that the first characteristic group is obtained by material division, and the second characteristic group is obtained by color division. In other preferred embodiments, the partitioning order may be flexibly adjusted according to influencing factors of a specific scenario.

After the feature data of each target object are obtained, it is also necessary to determine how many partition control can be performed by the light source module. In the six-ring eight-zone design of the LED ring-shaped cold light source of the light source module in this embodiment, it can theoretically perform independent control of eight zones, i.e. the number of zones is eight. It is also necessary to determine the calibration loop according to the height of the light source module and the workbench, the angle between the light and the workbench, etc., and the calibration loop should be matched with the target object, so as to ensure that the calibration loop can be located in the core area of the target object.

In another preferred embodiment, the calibration loop may also be determined according to the original shape of the target object, the main purpose of which is to calculate the integrated similarity and to divide the target object.

Referring to fig. 2, the target object is a gear (black area in the figure), and the calibration ring is a white ring according to the target object and the calibration ring set by the light source module, and the calibration ring coincides with the center of the gear reference circle. According to the number of the partitions corresponding to the light sources is eight, the target object can be divided into eight detection areas, and the target object can be divided into eight detection areas on average due to the regular and symmetrical overall shape of the gear. In fig. 2, the circle centers of the gear reference circles are considered to be divided, and the eight detection areas are respectively 1-2,2-3,3-4,4-5,5-6,6-7,7-8 and 8-1.

After the number of the marked wire loops and the subareas is determined, the marked wire loops are divided into a plurality of wire loop segments according to a set step length, each wire loop segment corresponds to one region in the target object, if the material characteristics and the color characteristics of two adjacent regions are similar, the regions corresponding to the adjacent wire loop segments can be combined, and the combined regions can be adjusted through the same light source parameters. And after the areas are combined, analyzing the comprehensive similarity between the areas and the next adjacent area until all the areas corresponding to the wire ring segments are judged to be finished, wherein the number of the finally formed areas is consistent with that of the subareas.

The calculation mode of the comprehensive similarity of the application can be referred as follows:

The similarity of materials is determined by the reflection characteristics of different materials on light mainly analyzing the target object. When the materials are consistent, the material similarity is high, and when the materials are inconsistent, the material similarity is also high when the reflection intensity of the light is basically the same. Color similarity is mainly evaluated by color difference, color space distance, color feature vector, color similarity index, and the like.

In this embodiment, in order to improve the calculation efficiency, only the comprehensive similarity on the adjacent wire loop segments can be calculated, so that the analysis of the areas corresponding to the adjacent wire loop segments is not needed, the data analysis amount can be reduced, and the method is suitable for target objects regularly distributed along the calibrated wire loops, such as the gear shown in fig. 2. Of course, the comprehensive similarity of the areas corresponding to the adjacent wire ring segments can be calculated, so that the influence of various factors in the areas is considered, and the merging reliability of the areas corresponding to the adjacent wire ring segments can be ensured.

However, when the difference between the region characteristic data corresponding to the adjacent wire loop segments is large, the comprehensive similarity of the regions corresponding to the adjacent wire loop segments should be evaluated. Referring to fig. 3, if color patches a, B and C exist in the region 8-1, and color patches B and C exist in the region 1-2 in fig. 3, the integrated similarity of the corresponding regions of the adjacent wire loop segments is evaluated. Taking the region 1-2 as an example, firstly calculating the comprehensive similarity of the two regions f-2 and e-f, combining the two regions f-2 and e-f to form a new region e-2 if the comprehensive similarity is larger than a similarity threshold, calculating the comprehensive i similarity of the region e-2 and d-e, and combining until the formed region (1-2) is not larger than the partition area after judging, thereby completing the division of one detection region.

After the comprehensive similarity of the adjacent wire loop segments (or corresponding regions) is calculated, it can be determined whether the comprehensive similarity is greater than a similarity threshold. When the integrated similarity is greater than the similarity threshold, it may be determined that the illumination parameters required by adjacent wire loop segments (or corresponding regions) corresponding to the integrated similarity are substantially identical (or not greatly different), and then the adjacent wire loop segments may be combined to form a new wire loop segment.

In another preferred embodiment, the segmentation analysis may be performed without the calibration loop, whether the integrated similarity of the adjacent setting frames is consistent may be determined according to the size of the setting frames (e.g., 1cm×1cm square), and a plurality of detection areas may be obtained by merging according to the determination result. Of course, the calculation method of the integrated similarity can be replaced by other possible methods, such as using ant colony, genetic algorithm, etc.

Before each merge, it should be verified whether the area after the merge will be larger than the partition area. When the area after the combination is not larger than the area of the partition, the combination treatment can be performed, but when the area of the combined area is larger than the area of the partition, the detection area formed after the combination cannot be subjected to illumination parameter adjustment through a single partition of the light source, so that the combination treatment is not performed.

It is noted that the final number of partitions in this embodiment is determined by the number of partitions of the light source. When the number of partitions is determined, the corresponding partition area is also determined. The determination principle of the detection areas is that the total number is not larger than the number of the subareas, and the comprehensive similarity of adjacent wire ring segments (or adjacent areas) in each detection area is closest, so that the subsequent unified adjustment of illumination parameters of each detection area is facilitated.

In another preferred embodiment, the light source is a six-ring eight-zone design, and can be configured as a plurality of detection zones at most, as shown in the schematic diagram of the clock detection zone in fig. 4. Eight zones can generate eight zones along the circumferential direction, 1-2 are one zone, and six zones can be generated along the radial direction by six rings, such as sector areas corresponding to the number a, b, c, d, e, f (not mutually included). In the subsequent illumination adjustment process, the sector areas corresponding to a, b, c, d, e, f can be used for independently realizing adjustment of illumination parameters. Region a, b, c, d, e, f is based on the effect after the integrated similarity merge.

It should be noted that the calibration wire loops in fig. 2, 3 and 4 are all circular, and are mainly determined according to the shape of the target object and the shape of the light source. In other embodiments, the calibration loop may be rectangular, only ensuring that it performs the zonal function of the detection zone.

Moreover, the main purpose of this embodiment is to ensure the quality of the acquired image data, i.e. to ensure that the edges and detail parts of the target object in the image quality are clear, so that it is not necessary to divide all the same areas into one detection area. For example, in fig. 2, the characteristic parameters corresponding to each detection area are consistent, and although the illumination parameters of each detection area can be theoretically controlled independently, the condition is only required to directly adjust the light source according to one illumination parameter.

After each detection zone is determined, the illumination parameters of each detection zone need to be acquired based on the illumination feature model. The illumination characteristic model in this embodiment is obtained by training an artificial intelligence model with a large amount of data. The standard training data for training comprises standard input data and standard output data, wherein the standard input data is a target object with various simulated materials and colors, and the labeling output data is an optimal illumination parameter corresponding to the standard input data. The optimal illumination parameters may be used to select a group of optimal imaging quality by continuously adjusting the illumination parameters. In other preferred embodiments, the existing data set may also be downloaded to complete training of the artificial intelligence model, which may be a BP neural network model or an RBF neural network model.

And extracting an illumination characteristic model, integrating and inputting the color characteristics and the material characteristics of each target object corresponding to the detection areas into the illumination characteristic model, and obtaining the optimal illumination parameters of each detection area. And carrying out partition adjustment on the light source in the light source module based on the illumination parameters.

The material adjustment can be performed by using a material model or a material number considered to be set, so that the material can be uniquely identified by the artificial intelligence model, and the color characteristics can be selected from at least one of dominant wavelength, chromaticity, brightness, color texture, color change rate and the like in the detection area, and the material is required to be input in a form identifiable by the artificial intelligence model.

It should be noted that, in order to ensure that the light source controlled by the partition can correspond to the detection area of the target object, a rotation mechanism may be provided on the workbench to adjust the position of the target object when the partition light source does not correspond to the detection area. Of course, the light source module can be rotatably arranged, and the position of each detection area corresponding to the target object is ensured by adjusting the light source module.

The method comprises the steps of obtaining a plurality of data, wherein part of data in the formula is obtained by removing dimensions and taking the numerical calculation, the formula is a formula closest to the actual situation by simulating a large amount of collected data through software, and preset parameters and preset thresholds in the formula are set by a person skilled in the art according to the actual situation or are obtained through simulating the large amount of data.

Working principle:

extracting characteristic data of each target object in the target group to be detected by combining the original shape, and associating the characteristic data with the detection sequence of the target objects to generate a characteristic sequence;

determining a target to be detected according to the characteristic sequence, and extracting characteristic data of the target to be detected;

and adjusting the light source according to the illumination parameters of the detection areas to further acquire image data of the target to be detected.

The above embodiments are only for illustrating the technical method of the present application and not for limiting the same, and it should be understood by those skilled in the art that the technical method of the present application may be modified or substituted without departing from the spirit and scope of the technical method of the present application.

Claims

1. The automatic imager detection system comprises a light source adjusting module and a data acquisition module connected with the light source adjusting module, and is characterized in that:

The method comprises the steps of constructing an illumination characteristic model, acquiring corresponding illumination parameters by combining characteristic data of a plurality of detection areas, adjusting light sources according to the illumination parameters of the plurality of detection areas, and further acquiring image data of an object to be detected, wherein the illumination characteristic model is constructed based on an artificial intelligent model;

partitioning the target to be detected based on the characteristic data and the light source characteristics, including:

Determining a plurality of detection areas based on the comprehensive similarity and the partition number;

The calculating the comprehensive similarity of the feature data of the object to be detected at the adjacent positions on the calibration wire loop comprises the following steps:

marking the material similarity and the color similarity of adjacent wire ring segments as CXD and YXD, calculating the comprehensive similarity ZXD through a formula ZXD=α1×CXD+α2×YXD, wherein α1 and α2 are weight coefficients, α1 is larger than or equal to α2, and the material similarity is used for evaluating whether the reflection characteristics of the materials are similar or not;

The determining a plurality of detection areas based on the comprehensive similarity and the partition number comprises the following steps:

2. The automated imager detection system of claim 1, wherein the extracting feature data of each object in the target set in combination with the original shape comprises:

3. The automated imager detection system of claim 1, wherein said acquiring corresponding illumination parameters in combination with the feature data of the plurality of detection zones comprises:

4. The automated imager detection system of claim 1, wherein said constructing an illumination feature model comprises:

The method comprises the steps of obtaining standard training data, wherein the standard training data comprises standard input data and standard output data, the standard input data is integrated based on material characteristics and color characteristics, and the standard output data is corresponding illumination parameters;

5. The automated imager detection system of claim 1, wherein said adjusting the light source according to the illumination parameters of the plurality of detection zones comprises: