An automated image analyzer detection system

Abstract

The present application discloses an automated imaging instrument detection system, which relates to the technical field of imaging instruments and solves the technical problems that the adjustment accuracy and efficiency of the light source module in the prior art are insufficient, affecting the imaging quality and detection efficiency of the target object detection; the present application partitions the target to be detected based on the feature data of the target object and the light source features to obtain several detection areas; determines the corresponding optimal lighting parameters according to the material features and color features of each detection area, and then completes the light source adjustment during the detection process; the present application adjusts the light source according to the feature data of each target object, which can improve the quality of the collected image data; the present application obtains a lighting feature model that can represent the mapping relationship between the material features, color features, and the corresponding optimal lighting parameters, and obtains the lighting parameters corresponding to each detection area through the lighting feature model; the present application obtains the lighting parameters of each detection area through the artificial intelligence model, which can improve the accuracy of light source adjustment and the image quality.

Description

Technical Field

[0001] This application belongs to the field of image instruments, relates to the automated detection technology of image instruments, and specifically is an image instrument detection system based on automation. Background Art

[0002] The fully automatic image measuring instrument is an artificial intelligence type modern optical non-contact measuring instrument developed on the basis of the digital image measuring instrument (also known as CNC image instrument). The fully automatic image measuring instrument inherits the excellent motion accuracy and motion control performance of the digital instrument, and integrates the design spirituality of the machine software, belonging to the current advanced optical dimension detection equipment.

[0003] When the existing image instrument detection system detects and analyzes a target object, multiple sets of hardware devices of the image instrument need to cooperate to collect the 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 process of collecting image data, and it is necessary to adaptively adjust the lighting parameters to ensure the imaging quality; but generally, the adjustment of the lighting parameters of the light source module is complex, which not only cannot ensure the imaging quality of different regions of the same target object, but also affects the detection efficiency of continuous different target objects.

[0004] This application provides an image instrument detection system based on automation to solve the above technical problems. Summary of the Invention

[0005] This application aims to solve at least one of the technical problems existing in the prior art; for this purpose, this application proposes an image instrument detection system based on automation, which is used to solve the technical problems that the adjustment accuracy and efficiency of the light source module in the prior art are insufficient, affecting the imaging quality and detection efficiency of target object detection.

[0006] To achieve the above object, the first aspect of this application provides an image instrument detection system based on automation, including: a light source adjustment module, and a data acquisition module connected thereto;

[0007] Data acquisition module: used to extract the feature data of each target object in the to-be-detected target group in combination with the original shape; associate the feature data with the detection order of the target object to generate a feature sequence; wherein, the feature data includes the original shape, and the material feature and color feature corresponding to each position in the original shape;

[0008] Light source adjustment module: used to determine the to-be-detected target according to the feature sequence, extract the feature data of the to-be-detected target; partition the to-be-detected target based on the feature data and the light source feature to obtain a number of detection areas; and,

[0009] Build a lighting feature model, obtain corresponding lighting parameters by combining the feature data of several detection areas; adjust the light source according to the lighting parameters of several detection areas, and then obtain the image data of the target to be detected; among them, the lighting feature model is built based on an artificial intelligence model.

[0010] When detecting a target object based on an imaging instrument, it mainly analyzes various parameters of the target object, such as size, shape, surface defects, etc. by collecting the image data of the target object and combining the built-in analysis algorithm. Therefore, the quality of the image data is crucial. However, during the image acquisition process, the configuration and adjustment of the light source will have a great impact on the image quality. There are usually many types of target objects detected by the imaging instrument, and it is difficult for the existing technology to automatically adjust the light source for different target objects, so it is difficult to obtain high-quality image data.

[0011] This application first determines the basic features of each target object in the group to be detected, and associates the basic features with the original shape of the target object to obtain feature data. Then, when detecting the target to be detected, according to its feature data and light source features, the target to be detected is divided into several detection areas. Then, the most suitable lighting parameters for each detection area are analyzed through the lighting feature model, and the light source is adjusted in zones during the detection process to ensure the quality of the collected image data.

[0012] Preferably, extract the feature data of each target object in the group of objects to be measured in combination with the original shape, including:

[0013] Obtain the basic features and original shape of the target object in the group of objects to be measured through the database; among them, the basic features include material features and color features;

[0014] First divide the original shape of the target object into several feature groups one according to the material features, and then divide the several feature groups one into several feature groups two according to the color features;

[0015] Associate the several feature groups two with their corresponding positions in the original shape to obtain feature data.

[0016] During the working process of the imaging instrument, the image quality will be affected by the material, material, etc. of the target object. If you want to improve the image quality by adjusting the light source in zones, you first need to divide the target object into zones. The basis for zoning is the factors that affect the image quality. Therefore, it is necessary to clarify the features of each position of the target object.

[0017] This application first extracts the basic features of each target object in the target group to be detected from the database, that is, the material features and color features of the target object. First, divide the original shape of the target object into multiple feature groups one according to the materials, and then divide the feature groups one into multiple feature groups two according to the color differences. Then each feature group two is obtained by dividing according to materials and colors. Associate each feature group two with its position in the original shape of the target object to obtain feature data.

[0018] After obtaining the feature data of the target to be detected, it is possible to quickly identify the material and color characteristics at each position of it, laying a data foundation for subsequently dividing the target object into multiple detection areas.

[0019] Preferably, partitioning the target to be detected based on the feature data and the light source features includes:

[0020] Extract the light source features of the light source module, and determine the number of partitions and the calibration line loop according to the light source features; wherein, the calibration line loop is determined according to the shape of the target object or the light source features;

[0021] Successively record the target objects in the feature sequence as the targets to be detected; calculate the comprehensive similarity of the feature data of the target to be detected at adjacent positions on the calibration line loop; wherein, the comprehensive similarity includes the material similarity and the color similarity;

[0022] Determine several detection areas based on the comprehensive similarity and the number of partitions.

[0023] When partitioning the target object, the first thing to consider is whether the light source module can be partitioned and controlled, and how many areas can be independently controlled; also consider what basis to partition the target object.

[0024] After this application determines that the light source module can be partitioned and controlled, it obtains the number of partitions. Determine the calibration line loop by the height between the light source module and the workbench and the angle between the light and the workbench. Of course, the size of the target object also needs to be considered when determining the calibration line loop. Of course, it can also be determined according to the shape of the target object and the light source features; for example, if the target object is circular, the calibration line loop can be set as circular, and if the target object is square, the calibration line loop can be set as square. The light source features include the light source shape and how many control areas the light source can be divided into; when the light source shape is circular, the calibration line loop can also be set as circular.

[0025] Calculate the comprehensive similarity of the feature data of the target to be detected at adjacent positions on the calibration line loop. Segment the calibration line loop based on the comprehensive similarity and the number of partitions, and then realize the partitioning of the target object. This application can ensure that the materials and colors in each detection area are similar, which is helpful for adjusting the light source responsible for this detection area, and can also ensure that the quality of the acquired image data after adjustment is good.

[0026] Preferably, calculating the comprehensive similarity of the feature data of the target to be detected at adjacent positions on the calibration line loop includes:

[0027] Dividing the calibration line loop into several line loop segments according to a set step size;

[0028] Mark the material similarity and color similarity of adjacent line loop segments as CXD and YXD; calculate the comprehensive similarity ZXD through the formula ZXD = α1×CXD + α2×YXD; where α1 and α2 are weight coefficients, and α1≥α2, and the material similarity is used to evaluate whether the material reflection characteristics are similar.

[0029] In this application, by segmenting the calibration line loop, several line loop segments are obtained, and then the comprehensive similarity of adjacent line loop segments is calculated. The calculation of the comprehensive similarity is mainly based on the feature data of the regions corresponding to each line loop segment, and it mainly judges whether the light source parameters required by adjacent line loop segments are similar based on materials, colors, etc. In this application, it is used to judge whether adjacent line loop segments are similar by calculating the feature data, so as to judge whether the requirements for light source parameters of adjacent line loop segments are similar. If they are similar, adjacent line loop segments can be merged.

[0030] Preferably, determining several detection areas based on the comprehensive similarity and the number of partitions includes:

[0031] Calculating and determining the partition area according to the number of partitions;

[0032] When the comprehensive similarity of adjacent line loop segments is greater than the similarity threshold, then judge whether the area where adjacent line loop segments are merged will exceed the partition area; if so, do not merge adjacent line loop segments; if not, merge adjacent line loop segments;

[0033] After adjusting the merging result, several detection areas are obtained.

[0034] After determining the number of partitions, 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, etc., and divide the target object into several detection areas through this partition area.

[0035] In this application, it is judged whether the comprehensive similarity is greater than the similarity threshold. If it is greater, the target object areas corresponding to adjacent line loop segments are merged, and it is judged whether this area is greater than the partition area; if it is greater than the partition area, no merging is performed, and if it is not greater than the partition area, adjacent line loop segments are merged to form a new line loop segment. Then, the new line loop segment is compared with the next adjacent line loop segment to judge whether merging is required.

[0036] Preferably, obtaining the corresponding illumination parameters by combining the feature data of several detection areas includes:

[0037] Successively extracting the material feature and color feature from the feature data of the detection area;

[0038] Integrate the material features and color features into the model input data of the light feature model; input the model input data into the light feature model to obtain the light parameters of the detection area; wherein, the light parameters include light intensity and color temperature.

[0039] It is difficult to express the relationship between the material features, color features and light parameters with a very clear mapping relationship. Therefore, this application uses an artificial intelligence model with strong non-linear fitting ability to achieve. Extract the material features and color features from the feature data of each detection area, and input the integrated two into the constructed light adjustment model to obtain the corresponding light parameters of the detection area.

[0040] Preferably, constructing the light feature model includes:

[0041] Obtain standard training data; wherein, the standard training data includes standard input data and standard output data. The standard input data is integrated based on the material features and color features, and the standard input data is the corresponding light parameters;

[0042] Train the artificial intelligence model with the standard training data, and mark the trained artificial intelligence model as the light feature model; wherein, the standard training data is obtained through a large number of scene simulations.

[0043] When training the artificial intelligence model in this application, it is necessary to conduct simulation experiments on different material features and color features of the detection area to obtain their suitable light parameters. After obtaining multiple groups of data through simulation, the multiple groups of data sets can be preprocessed to obtain standard training data. The preprocessing includes abnormal data elimination, data augmentation, etc. Then, the artificial intelligence model can be trained with the standard training data to obtain the light feature model.

[0044] Preferably, adjusting the light source according to the light parameters of several detection areas includes:

[0045] Extract several detection areas of the target to be detected and the corresponding light parameters;

[0046] When the target to be detected enters the image acquisition area, adjust the light source according to the light parameters of each detection area; after the adjustment is completed, collect the image data of the target to be detected.

[0047] After obtaining several detection areas of each target to be detected and the corresponding light parameters, once the target to be detected enters the image acquisition area (set on the workbench), the light source can be adjusted in zones based on the light parameters, which can ensure that the light of each detection area of the target to be detected can ensure high-quality image data is collected.

[0048] Compared with the prior art, the beneficial effects of this application are:

[0049] 1. This application extracts each target object from the target group to be detected in sequence, partitions the target to be detected based on the feature data of the target object and the light source features, and obtains several detection areas; then determines the corresponding optimal lighting parameters according to the material features and color features of each detection area, and further completes the light source adjustment during the detection process; this application adjusts the light source according to the feature data of each target object, which can improve the quality of the collected image data.

[0050] 2. This application trains an artificial intelligence model with a large amount of standard training data to obtain a lighting feature model that can represent the mapping relationship between material features, color features, and the corresponding optimal lighting parameters, and obtains the lighting parameters corresponding to each detection area through the lighting feature model; this application obtains the lighting parameters of each detection area through the artificial intelligence model, which can improve the accuracy of light source adjustment and thus improve the image quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0052] Figure 1 It is a schematic diagram of the system principle of the embodiment of the present application;

[0053] Figure 2 It is a schematic diagram of the gear detection area of the embodiment of the present application Figure 1 ;

[0054] Figure 3 It is a schematic diagram of the gear detection area of the embodiment of the present application Figure 2 ;

[0055] Figure 4 It is a schematic diagram of the clock detection area of the embodiment of the present application;

[0056] Figure 5 It is a schematic diagram of the method flow of the embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] The following will clearly and completely describe the technical solutions of the present application in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0058] Please refer to Figures 1-5, an embodiment of the first aspect of the present application provides an automated imaging instrument detection system, including: a light source adjustment module and a data acquisition module connected thereto;

[0059] Data acquisition module: used to extract the feature data of each target object in the target object group to be measured in combination with the original shape; associate the feature data with the detection order of the target object to generate a feature sequence; wherein, the feature data includes the original shape, as well as the material features and color features corresponding to each position in the original shape;

[0060] Light source adjustment module: used to determine the target to be detected according to the feature sequence, extract the feature data of the target to be detected; partition the target to be detected based on the feature data and the light source features to obtain several detection areas; and,

[0061] Construct a lighting feature model, obtain the corresponding lighting parameters by combining the feature data of several detection areas; adjust the light source according to the lighting parameters of several detection areas, and then obtain the image data of the target to be detected; wherein, the lighting feature model is constructed based on an artificial intelligence model.

[0062] The imaging instrument of the present application mainly includes the following hardware:

[0063] Imaging module: It uses a high-resolution color CCD with an imported original Japanese Sony chip, has excellent imaging, is clear and vivid, and has rich colors. The optical lens uses imported high-quality optical components, undergoes multi-layer optical coating, is resistant to oil stains and corrosion, has a small optical attenuation, and remains bright and clear after long-term use.

[0064] Light source module: The imported LED ring cold light source has a six-ring and eight-zone design, stepless adjustment, delicate and soft light, and bright and clear surface light imaging. The imaging is clear and the edges and corners are distinct, especially for the measurement of shaft-like and arc-shaped workpieces, which is more accurate. The light source adjustment module mainly adjusts and controls the light source of the light source module.

[0065] Drive module: CNC full-closed loop controller (four axes), high-precision full-closed loop servo motor, integrated C3-level double angular contact bearings, ground C3-level ball screws. Full-closed loop control, high-speed response, accurate positioning, and stable operation.

[0066] Lubrication module: oil circuit distributor, screw nut, guide rail slider oil injector, high-pressure oil pipe, special oil for THK guide rails and screws in Japan. The standardized oil circuit design and high-quality grease ensure that the instrument can maintain long-term accuracy and operation stability during long-term use and high-speed operation.

[0067] Vibration damping module: The configuration of the vibration damper enables the instrument to obtain higher accuracy, speed and stability in a vibrating use environment or at a high-speed operation state.

[0068] Glass stage: The original glass is imported from SCHOTT in Germany and manufactured by the micro-float process. After precision grinding, it has extremely high flatness and light transmittance.

[0069] Workbench: The three-layer workbench is designed with high-strength aviation aluminum alloy material. Through special treatment processes, the interior and surface of the material have super hardness and stability. The long-term natural aging process eliminates the change of internal stress in the workbench, and its physical properties are very stable.

[0070] Base column: The base and column are made of high-precision natural granite and processed by 00-level grinding. Their physical properties are stable and there is no change in internal stress, enabling the three axes to have the same temperature characteristics and expansion coefficients, ensuring the long-term accuracy, reliability, and stability of the instrument.

[0071] Before adjusting the light source, it is necessary to clarify which target objects are to be detected and what characteristics of these target objects affect the light source adjustment.

[0072] In this application, the target group to be detected is first extracted from the database. The target group to be detected includes several target objects that need to be detected by the imaging instrument, as well as the basic characteristics and original shapes of each target object. Here, the original shape is the contour shape of the target object. For example, the original shape of a gear is a round cake shape; the basic characteristics are some characteristics that affect the light source adjustment. For example, the reflection of light by the material will affect the image quality, and the color of the object will also affect the image quality. Therefore, the target objects in the target group to be detected are the detection objects, and the basic characteristics of each target object are the key to affecting the light adjustment.

[0073] After clarifying the above content, the original shapes of the target objects are first divided according to the material characteristics, that is, the areas with the same material are marked in the original shape, and each mark is used as a feature group one; in each feature group one, it is divided according to the color characteristics to obtain feature group two. At least one feature (material feature or color feature) of each position corresponding to each feature group two is different from the corresponding positions of other feature groups two.

[0074] The divided feature group two is associated with its position in the original shape, that is, each position in the original shape is associated with a feature group two. In principle, the lighting parameters of each position corresponding to each feature group two should be adjusted separately. However, considering the number of partitions of the light source module and the influence of light source interference, this application also needs to merge the positions in the original shape of the target object to obtain several detection areas.

[0075] It should be noted that the image data collected by the imager in this application is default used to analyze the size, defects, etc. of the target object. Here, the reflection intensity of the material has a greater impact. Therefore, the first feature group is obtained by material division, and then the second feature group is obtained by color division. In some other preferred embodiments, the division order can be flexibly adjusted according to the influencing factors of the specific scenario.

[0076] After obtaining the feature data of each target object, it is also necessary to determine how many partition controls the light source module can perform. The light source module in this embodiment is a six-ring and eight-zone design of an LED ring cold light source. In theory, it can perform independent control of eight zones, that is, the number of partitions is eight. It is also necessary to determine the calibration line ring according to the height of the light source module from the workbench, the angle of the light with the workbench, etc. The calibration line ring should be matched with the target object to ensure that the calibration line ring can be in the core area of the target object.

[0077] In another preferred embodiment, the calibration line ring can also be determined according to the original shape of the target object, and its main purpose is to calculate the comprehensive similarity and divide the target object.

[0078] For the determination of the calibration line ring, please refer to Figure 2 , the target object is a gear (the black area in the figure). According to the target object and the light source module, the set calibration line ring is a white circular ring, and the center of the calibration line ring coincides with the center of the gear pitch circle. According to the number of partitions corresponding to the light source being eight, the target object can be divided into eight detection zones. Since the overall shape of the gear is regular and symmetric, the target object can be evenly divided into eight detection zones. Figure 2 It is considered to divide with the center of the gear pitch circle, and the eight detection zones obtained are 1-2, 2-3, 3-4, 4-5, 5-6, 6-7, 7-8, 8-1.

[0079] After determining the calibration line ring and the number of partitions, the calibration line ring is divided into several line ring segments according to the set step size. Each line ring segment corresponds to a region in the target object. If the material features and color features of two adjacent regions are similar, the regions corresponding to the adjacent line ring segments can be merged, and the merged region can be adjusted by the same light source parameters. After the regions are merged, the region is regarded as a whole, and its comprehensive similarity with the next adjacent region is analyzed until all the regions corresponding to the line ring segments are judged. Finally, the number of formed regions is the same as the number of partitions.

[0080] The calculation method of the comprehensive similarity of this application can be referred to as follows:

[0081] Mark the material similarity and color similarity of adjacent line loop segments as CXD and YXD; calculate the comprehensive similarity ZXD through the formula ZXD = α1×CXD + α2×YXD; where α1 and α2 are weight coefficients, and α1 ≥ α2, and the material similarity is used to evaluate whether the material reflection characteristics are similar.

[0082] The material similarity is mainly analyzed by the reflection characteristics of light by different materials of the target object. When the materials are the same, the material similarity is high; when the materials are different and the reflection intensities of light are basically the same, the material similarity is also relatively high. The color similarity is mainly evaluated by color difference, color space distance, color feature vector, color similarity index, etc.

[0083] In this embodiment, in order to improve the calculation efficiency, the comprehensive similarity on adjacent line loop segments can be calculated only, so that it is not necessary to analyze the regions corresponding to adjacent line loop segments, which can reduce the amount of data analysis and is applicable to target objects regularly distributed along the calibrated line loop, such as Figure 2 the shown gear. Of course, the comprehensive similarity of the regions corresponding to adjacent line loop segments can also be calculated, which takes into account the influence of various factors within the region and can ensure the reliability of the merger of the regions corresponding to adjacent line loop segments.

[0084] However, when the regional feature data corresponding to adjacent line loop segments differ greatly, the comprehensive similarity of the regions corresponding to adjacent line loop segments should be evaluated. Please refer to Figure 3 , Figure 3 In, there is a color block A in region 8-1 and color blocks B and C in 1-2, then the comprehensive similarity of the regions corresponding to adjacent line loop segments is evaluated. Taking region 1-2 as an example, first calculate the comprehensive similarity of the two regions f-2 and e-f. If the comprehensive similarity is greater than the similarity threshold, then merge the two regions f-2 and e-f to form a new region e-2, calculate the comprehensive similarity of region e-2 and d-e, and judge and merge until the formed region (1-2) is not larger than the partition area, then the division of a detection area is completed.

[0085] After calculating the comprehensive similarity of adjacent line loop segments (or corresponding regions), it can be judged whether the comprehensive similarity is greater than the similarity threshold. When the comprehensive similarity is greater than the similarity threshold, it can be determined that the illumination parameters required for the adjacent line loop segments (or corresponding regions) corresponding to the comprehensive similarity are basically the same (or not much different), and then the adjacent line loop segments can be merged to form a new line loop segment.

[0086] In another preferred embodiment, instead of performing segmented analysis according to the calibrated line loops, the comprehensive similarity of adjacent set frames (such as a 1 cm × 1 cm square) can be judged to determine whether they are the same, and several detection areas can be obtained by merging according to the judgment results. Of course, the calculation method of the comprehensive similarity can also be replaced by other possible methods, such as using algorithms like ant colony and genetic algorithms.

[0087] Before each merge, it should be verified whether the area of the merged region will be larger than the partition area. When the area of the merged region is not larger than the partition area, the merge process can be carried out; but when the area of the merged region is larger than the partition area, the detection area formed after the merge cannot be adjusted for the lighting parameters through a single partition of the light source, so no merge process is performed.

[0088] It should be 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 principle for determining the detection areas is that the total number is not greater than the number of partitions, and the comprehensive similarity of adjacent line loop segments (or adjacent regions) in each detection area is the closest, which is helpful for subsequent unified adjustment of the lighting parameters for each detection area.

[0089] In another preferred embodiment, when the light source is designed with six rings and eight zones, the maximum number of detection areas that can be set is multiple, such as Figure 4 as shown in the schematic diagram of the clock detection area. The eight zones can generate eight partitions along the circumference, and 1-2 is one of the partitions; the six rings can generate six partitions along the radial direction, such as the fan-shaped regions corresponding to the numbers a, b, c, d, e, f (not including each other). During the subsequent lighting adjustment process, the fan-shaped regions corresponding to a, b, c, d, e, f can all independently adjust the lighting parameters. The regions a, b, c, d, e, f are the effects after merging based on the comprehensive similarity.

[0090] It should be noted that Figure 2 、 Figure 3 and Figure 4 the calibrated line loops in are all circular, which are mainly determined according to the shape of the target object and the shape of the light source. In some other embodiments, the calibrated line loops can be rectangular, as long as they can achieve the partition function of the detection area.

[0091] Moreover, the main purpose of this embodiment is to ensure the quality of the collected image data, that is, to ensure that the edges and details of the target object in the image quality are clear, so it is not overly required to divide all the same regions into one detection area. For example Figure 2 the characteristic parameters corresponding to each detection area are the same. Although theoretically the lighting parameters of each detection area can be controlled separately, in this case, the light source can be adjusted according to one lighting parameter directly.

[0092] After determining each detection area, it is necessary to obtain the lighting parameters of each detection area based on the lighting feature model. The lighting feature model in this embodiment is obtained by training an artificial intelligence model with a large amount of data. The standard training data for training includes standard input data and standard output data. The standard input data is target objects of various materials and colors simulated, and the labeled output data is the optimal lighting parameters corresponding to the standard input data. The optimal lighting parameters can be selected by continuously adjusting the lighting parameters to select a set with the optimal imaging quality. In some other preferred embodiments, existing data sets can also be downloaded to complete the training of the artificial intelligence model. The artificial intelligence model can be a BP neural network model or an RBF neural network model.

[0093] Extract the lighting feature model, integrate the color features and material features of the corresponding detection areas of each target object and input them into the lighting feature model, and the optimal lighting parameters of each detection area can be obtained. Based on this lighting parameter, the light sources in the light source module can be adjusted in zones.

[0094] It should be noted that for material adjustment, the material model number or a set material number can be used, as long as it is unique and can be recognized by the artificial intelligence model; for color features, at least one of the dominant wavelength, chromaticity, brightness, color texture, color change rate, etc. in the detection area can be selected, and it needs to be input in a form that can be recognized by the artificial intelligence model.

[0095] It is worth noting that in order to ensure that the partition-controlled light source can correspond to the detection area of the target object, a rotating mechanism can be set 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 set to be rotatable, and the position of the light source module can be adjusted to ensure correspondence with the detection areas of each target object.

[0096] Some of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is obtained by software simulation of a large amount of collected data to get a formula closest to the actual situation; the preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained by simulation of a large amount of data.

[0097] Working principle:

[0098] Combine the original shape to extract the feature data of each target object in the target group to be measured; associate the feature data with the detection order of the target object to generate a feature sequence;

[0099] Determine the target to be detected according to the feature sequence, extract the feature data of the target to be detected; partition the target to be detected based on the feature data and the light source features to obtain several detection areas;

[0100] Construct a lighting feature model, obtain corresponding lighting parameters by combining the feature data of several detection areas; adjust the light source according to the lighting parameters of several detection areas, and then obtain the image data of the target to be detected.

[0101] The above embodiments are only used to illustrate the technical method of the present application rather than to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present application can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present application.

Claims

1. An automated imaging instrument detection system, comprising: The light source adjustment module and the data acquisition module connected thereto are characterized in that: Data acquisition module: used to extract characteristic data of each target object in the target group to be tested in combination with the original shape; associate the characteristic data with the detection order of the target object to generate a characteristic sequence; wherein the characteristic data includes the original shape, and the material characteristics and color characteristics corresponding to each position in the original shape; Light source adjustment module: used to determine the target to be detected according to the feature sequence, extract the feature data of the target to be detected; partition the target to be detected based on the feature data and the light source characteristics to obtain a number of detection areas; and, Constructing an illumination feature model, and obtaining corresponding illumination parameters in combination with the feature data of several detection areas; adjusting the light source according to the illumination parameters of several detection areas, and then obtaining image data of the target to be detected; wherein the illumination feature model is constructed based on an artificial intelligence model; The partitioning of the target to be detected based on the feature data and the light source feature includes: Extracting the light source characteristics of the light source module, and determining the number of partitions and the calibration line ring according to the light source characteristics; wherein the calibration line ring is determined according to the shape of the target object or the light source characteristics; The target objects in the feature sequence are recorded as the targets to be detected in turn; the comprehensive similarity of the feature data of the targets to be detected at adjacent positions on the calibration line loop is calculated; wherein the comprehensive similarity includes material similarity and color similarity; Determine a number of detection zones based on the comprehensive similarity and the number of zones; The step of calculating the comprehensive similarity of the feature data of the target to be detected at adjacent positions on the calibration line loop includes: Divide the calibration line loop into a number of line loop segments according to the set step length; The material similarity and color similarity of adjacent line loop segments are marked as CXD and YXD; the comprehensive similarity ZXD is calculated by the formula ZXD=α1×CXD+α2×YXD; where α1 and α2 are weight coefficients, and α1≥α2, and the material similarity is used to evaluate whether the material reflection characteristics are similar; The method of determining a plurality of detection zones based on the comprehensive similarity and the number of partitions includes: Determine the partition area based on the number of partitions; When the comprehensive similarity of adjacent line loop segments is greater than the similarity threshold, it is determined whether the area to be merged of the adjacent line loop segments exceeds the partition area; if yes, the adjacent line loop segments are not merged; if no, the adjacent line loop segments are merged; After adjusting the combined result, several detection areas are obtained.

2. The automated imaging system according to claim 1, characterized in that: The extracting of feature data of each target object in the target group to be measured in combination with the original shape includes: Obtaining basic features and original shapes of target objects in the target group to be tested through a database; wherein the basic features include material features and color features; The original shape of the target object is first divided into a number of feature groups 1 according to the material feature, and then the number of feature groups 1 are divided into a number of feature groups 2 according to the color feature; The plurality of feature groups are associated with their corresponding positions in the original shape to obtain feature data.

3. The automated imaging system according to claim 1, characterized in that: The step of combining the characteristic data of a plurality of detection areas to obtain corresponding illumination parameters includes: extracting material features and color features from the feature data of the detection area in sequence; The material characteristics and color characteristics are integrated into model input data of the illumination characteristic model; the model input data is input into the illumination characteristic model to obtain illumination parameters of the detection area; wherein the illumination parameters include illumination intensity and color temperature.

4. The automated imaging system according to claim 1, characterized in that: The step of constructing the illumination feature model comprises: Obtaining standard training data; wherein the standard training data includes 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 the corresponding lighting parameters; The artificial intelligence model is trained through standard training data, and the trained artificial intelligence model is marked as a lighting feature model; wherein the standard training data is obtained through a large number of scene simulations.

5. The automated imaging detection system according to claim 1, characterized in that: The step of adjusting the light source according to the illumination parameters of the plurality of detection zones includes: Extracting several detection areas of the target to be detected and corresponding illumination parameters; When the target to be detected enters the image acquisition area, the light source is adjusted according to the illumination parameters of each detection area; after the adjustment is completed, the image data of the target to be detected is collected.

Images