Large component appearance visual inspection apparatus

Abstract

The application discloses a large-sized part appearance visual detection equipment, and relates to the field of laser measurement technology.The equipment comprises a laser emission and receiving module, a laser measurement control unit and a data fusion and processing module.The laser measurement control unit is used for processing and analyzing the received laser signals to obtain distance information of the part surface and a three-dimensional model of the actual part.The data fusion and processing module is used for fusing and processing the visual detection data and the laser measurement data, analyzing and processing the fused data through a specific algorithm, judging whether the whole part and each part is qualified or not, and generating a detection report.The application can realize accurate detection of small defects on the surface of the large-sized part and accurate size measurement by introducing the laser measurement technology, and further improves the detection accuracy by organically fusing the visual detection data.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of laser measurement technology, and in particular to a large part of visual appearance detection equipment. BACKGROUND

[0002] With the development of modern industry, large parts are more and more widely used in the fields of aerospace, automobile manufacturing, new energy and so on, and the quality and performance requirements of large parts are also higher and higher, so more accurate and efficient detection methods are needed, and multiple types of auxiliary detection can further improve the detection accuracy.

[0003] The existing multi-probe detection device often uses a spherical fixed support to install and fix multiple detection probes, and when multiple probes are operated, the multiple detection probes cannot be adjusted in a narrow area due to the fixed spacing, thereby causing inconvenience for detection, and the multiple detection probes installed by the spherical fixed support need to be rotated by the spherical fixed support to repeatedly detect a single area, which reduces the detection efficiency. Therefore, the present application designs a large part of visual appearance detection equipment to solve the above problems. SUMMARY

[0004] The purpose of the present application is to solve the problems existing in the prior art and provide a large part of visual appearance detection equipment.

[0005] In order to achieve the above purpose, the present application adopts the following technical scheme: a large part of visual appearance detection equipment, comprising a bottom plate, characterized in that the top surface of the bottom plate is provided with a placing box, the side end of the placing box is provided with a sliding guide seat, the sliding guide seat is vertically arranged on the bottom plate, a mechanical hand detection mechanism is slidably arranged on the sliding guide seat, the mechanical hand detection mechanism is composed of a mechanical gripper and a detection assembly, the mechanical gripper is slidably sleeved on the sliding guide seat on one side, the other side of the mechanical gripper is provided with a butt joint seat, the detection assembly comprises a detection main column, a plurality of movable sleeves, a rotary motor and a visual detector; the visual detector is used for collecting visual detection data of the part;

[0006] It also includes a laser measurement detection assembly, which includes a laser emission and reception module, a laser measurement control unit, a data fusion and processing module;

[0007] The laser emission and reception module is used for sending laser signals and receiving laser signals reflected from the surface of the part;

[0008] The laser measurement control unit is used for adjusting the laser emission parameters according to the detection requirements; processing and analyzing the received laser signals to obtain the distance information of the part surface and the three-dimensional model of the actual part; wherein the laser emission parameters include the frequency, power and emission angle of the laser emission.

[0009] The data fusion and processing module is used for fusing the visual detection data and the laser measurement data, analyzing and processing the fused data through a specific algorithm, judging whether the whole and each part of the part is qualified, and generating a detection report.

[0010] Preferably, the bottom end of each support column of the placing box is provided with a protective support pad, and the upper end of the placing box is provided with a support frame for supporting large parts.

[0011] Preferably, the detection assembly is composed of one detection main column and a plurality of movable sleeves, the movable sleeves are equidistantly clamped on the detection main column, a rotary motor is equidistantly installed in the detection main column, a transmission disc is installed on the motor shaft of the rotary motor, protrusions are equidistantly protruded on the transmission disc, a fan-shaped notch for rotating the protrusions is formed in the detection main column, the protrusions of the transmission disc are fixedly connected with the movable sleeves, and a visual detector is arranged at the bottom end of the detection main column.

[0012] Preferably, a sliding detection column is slidably arranged in each movable sleeve, a detection probe is installed at the bottom end of the sliding detection column, two sliding gears driven by micro-motors are arranged at the upper end of the sliding detection column, a sliding tooth groove for cooperating with the sliding gears is formed in the movable sleeve, and a limiting sliding groove for clamping the sliding detection column is formed in the movable sleeve.

[0013] Preferably, a plurality of groups of laser emission and receiving modules are arranged, each group of laser emission and receiving modules is installed at the position of each movable sleeve corresponding to the detection main column, a laser beam is emitted by the laser emission device and irradiated on the surface of the large part, and the laser receiving device receives the laser signal reflected from the surface of the part.

[0014] Preferably, the received laser signal is processed and analyzed, specifically:

[0015] The detection requirements of the part are identified, including the material, surface characteristics and detection accuracy requirements of the large part; the frequency, power and emission angle of the laser emission are determined according to the detection requirements;

[0016] The length of time from laser emission to reception is obtained, and the distance from the laser emission point to the measurement point on the surface of the part is calculated as a measurement distance according to the propagation speed of laser in air by using a preset formula;

[0017] A three-dimensional model of the actual part is constructed by the principle of triangulation combined with the emission angle and the relative position relationship of the measurement points.

[0018] Preferably, the execution steps of the data fusion and processing module are specifically:

[0019] Data alignment: a feature point-based matching algorithm is adopted to process the laser measurement data and the visual detection data, feature points are extracted from the visual images in the visual detection data and the laser measurement data, and the feature points are matched by using their descriptors to determine the correspondence relationship between the two, so as to realize the alignment of the data in space;

[0020] Constructing a standard three-dimensional model and judging defects: a standard three-dimensional model of the part is preset according to the design drawing or standard specification of the part; wherein the standard three-dimensional model defines the size, shape and factory range information of each part of the part;

[0021] Comparing and analyzing the standard three-dimensional model with the three-dimensional model of the actual part, determining whether there is a protruding or recessed area by calculating the deviation of each point between the standard three-dimensional model and the three-dimensional model of the actual part; identifying the area and depth of the protrusion or recess in the protruding or recessed area, and marking the area as a potential defect area when the area or depth of the protrusion or recess exceeds the preset tolerance range; and calculating the area and depth of the protrusion or recess in the protruding or recessed area to obtain a defect score;

[0022] Dividing the three-dimensional model of the actual part into a plurality of regions according to the structural characteristics and functional importance of the part, and setting a region importance weight for each region;

[0023] Identifying the defect score of the potential defect area in each region and the total number of defect areas;

[0024] Weighting all potential defect areas in the region and the total number of defect areas to obtain a region defect value;

[0025] Then, weighting the region defect values of all divided regions of the part with their corresponding region importance weights to obtain a comprehensive defect score of the part;

[0026] If the comprehensive defect value is greater than the preset defect threshold value, the part is unqualified; otherwise, the part is qualified;

[0027] Finally, generating a detection report: marking the basic information of the part, the qualification rate in the preset production cycle, the change of qualification trend, the detailed information of all potential defect areas, the defect situation of each region and the overall comprehensive defect score as a detection report; the detection report is presented in a combination of visual charts and textual descriptions.

[0028] As preferred, the data fusion and processing module further comprises a marking display module;

[0029] The mark display module is used for matching the defect score of each defect on the spare part with a preset defect mark value range group when the spare part is unqualified, so as to obtain the corresponding defect mark; and the defect mark is visually marked and displayed on the preset three-dimensional model of the spare part.

[0030] Compared with the prior art, the large spare part appearance visual detection equipment provided by the application has the following beneficial effects:

[0031] 1、The application can realize accurate detection of small defects on the surface of large spare parts and accurate size measurement by introducing laser measurement technology, and further improve the detection accuracy by organic integration with visual detection data.

[0032] 2、The application can quickly adapt to the complex structure of the spare part and the needs of different detection parts by flexibly adjusting the movable sleeve and the sliding detection column in the detection assembly of the mechanical hand detection mechanism; at the same time, the quick movement of the mechanical gripper makes the detection process more efficient, and the large spare part can be detected in all directions and at multiple angles, effectively improving the detection efficiency and shortening the detection cycle. BRIEF DESCRIPTION OF DRAWINGS

[0033] The drawings described herein are used to provide further understanding of the application, and form a part of the application. The schematic embodiments of the application and their descriptions are used to explain the application, and do not constitute an improper limitation on the application. In the drawings:

[0034] Figure 1 It is a whole schematic view of the large spare part appearance visual detection equipment proposed by the application;

[0035] Figure 2 It is a perspective view of the placing box proposed by the application;

[0036] Figure 3 It is a perspective view of the mechanical hand detection assembly proposed by the application;

[0037] Figure 4 It is a perspective view of the detection assembly proposed by the application;

[0038] Figure 5 It is a principle block diagram of the laser measurement detection assembly proposed by the application.

[0039] In the drawings, the serial numbers are as follows: 1, bottom plate; 2, placing box; 3, sliding guide seat; 4, mechanical hand detection mechanism; 5, protection support pad; 6, support frame; 7, mechanical gripper; 8, butt joint seat; 9, detection main column; 10, rotary motor; 11, sliding detection column; 12, sliding gear slot; 13, sliding gear; 14, limit sliding groove. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application.

[0041] Referring to Figures 1-5 The large part appearance visual detection device in the present application comprises a bottom plate 1, the top surface of the bottom plate 1 is provided with a placing box 2, the side end of the placing box 2 is provided with a sliding guide seat 3, the sliding guide seat 3 is vertically arranged on the bottom plate 1, a mechanical hand detection mechanism 4 is slidingly arranged on the sliding guide seat 3, the mechanical hand detection mechanism 4 is composed of a mechanical gripper 7 and a detection assembly, the mechanical gripper 7 is slidingly sleeved on the sliding guide seat 3 on one side, the other side of the mechanical gripper 7 is provided with a butt joint seat 8, the detection assembly comprises a detection main column 9, a plurality of movable sleeves, a rotating motor 10 and a visual detector; the visual detector is used for collecting visual detection data of the part;

[0042] The laser measurement detection assembly comprises a laser emission and receiving module, a laser measurement control unit and a data fusion and processing module;

[0043] The laser emission and receiving module is used for sending a laser signal and receiving a laser signal reflected from the surface of the part;

[0044] The laser measurement control unit is used for adjusting laser emission parameters according to detection requirements; processing and analyzing the received laser signal to obtain distance information of the surface of the part and a three-dimensional model of the actual part; wherein the laser emission parameters comprise frequency, power and emission angle of the laser emission;

[0045] The data fusion and processing module is used for fusing and processing the visual detection data and the laser measurement data, analyzing and processing the fused data through a specific algorithm, judging whether the whole part and each part are qualified, and generating a detection report.

[0046] In the present application, the bottom end of the support column of the placing box 2 is provided with a protective support pad 5, and the upper end of the placing box 2 is provided with a support frame 6 for supporting the large part.

[0047] In the present application, the detection assembly is composed of a detection main column 9 and a plurality of movable sleeves, the plurality of movable sleeves are equidistantly clamped on the detection main column 9, the detection main column 9 is equidistantly provided with a rotating motor 10, a transmission disc is installed on the motor shaft of the rotating motor 10, a protrusion is equidistantly protruded on the transmission disc, a fan-shaped notch for rotating the protrusion is formed on the detection main column 9, the protrusion of the transmission disc is fixedly connected with the movable sleeve, and the bottom end of the detection main column 9 is provided with a visual detector.

[0048] In the application, the sliding detection column 11 is arranged in the movable sleeve, the bottom end of the sliding detection column 11 is provided with a detection probe, the upper end of the sliding detection column 11 is provided with two sliding gear wheels 13 driven by micro motors, the movable sleeve is provided with a sliding tooth groove 12 for matching the sliding gear wheels 13, and the movable sleeve is provided with a limiting sliding groove 14 for clamping the sliding detection column 11.

[0049] In the application, the laser emission and receiving module is provided with a plurality of groups, each group of laser emission and receiving module is arranged at the position of the detection main column 9 corresponding to each movable sleeve, the laser emission device emits a laser beam, which irradiates on the surface of the large part, and the laser receiving device receives the laser signal reflected from the surface of the part; the laser emission and receiving module comprises a laser emission device and a laser emission device, which are arranged at different positions, so that the surface of the part can be measured from different angles, so that more comprehensive three-dimensional information can be obtained.

[0050] In the application, the received laser signal is processed and analyzed, specifically:

[0051] The detection requirements of the part are identified, including the material, surface characteristics and detection accuracy requirements of the large part; the frequency f, power P and emission angle θ of the laser emission are determined according to the detection requirements;

[0052] The length of time from laser emission to reception is recorded as t, according to the propagation speed c of laser in air, and the preset formula The distance from the laser emission point to the part surface measurement point is calculated as the measurement distance d; the measurement distance of the position is repeatedly obtained as di (i=1, 2,..., n), and the measurement distance is pretreated, including data filtering and outlier rejection processing; the pretreated measurement distance is recorded as the pretreatment measurement distance , i represents the number of any value in a group of different measurement distances;

[0053] The three-dimensional model of the actual part is constructed by the triangulation principle combined with the emission angle and the relative position relationship of the measurement point, specifically:

[0054] The position of laser emission is taken as the origin O, the horizontal direction is taken as the x axis, the vertical direction is taken as the y axis, and the direction perpendicular to the x and y axes is taken as the z axis;

[0055] According to the triangulation principle combined with the emission angle combined with the pretreatment measurement distance , the three-dimensional coordinates of each measurement point in the coordinate system are calculated ; any three-dimensional coordinate formula is , , ;

[0056] The three-dimensional coordinates of all the measuring points are combined to form point cloud data; the point cloud data of multiple angles and positions is obtained to obtain multiple sets of point cloud data; the multiple sets of point cloud data are registered by using an iterative closest point algorithm, so that the distance between each set of point cloud is minimized; the point cloud is divided into a plurality of voxel grids by using voxel grid filtering, and only one representative point is retained in each voxel grid;

[0057] The processed point cloud data is converted into a continuous surface model by a surface reconstruction algorithm (for example, a moving least square method is used, a local surface is fitted in the neighborhood of each point, and a surface model of the entire part surface is obtained by continuously moving the neighborhood); the reconstructed surface model is optimized, thereby constructing a three-dimensional model of the actual part; the optimization includes smoothing and hole filling; when constructing the three-dimensional model of the actual part, an octree-based point cloud compression algorithm can also be used to compress the processed point cloud data to reduce data storage and processing time.

[0058] In the present application, the execution steps of the data fusion and processing module are as follows:

[0059] Data alignment: a feature point-based matching algorithm (such as the SIFT algorithm) is used to process the laser measurement data and the visual detection data, feature points are extracted from the visual image in the visual detection data and the laser measurement data, and the feature points are matched by using their descriptors to determine their corresponding relationship and achieve spatial alignment of the data, which is known in the art and will not be described in detail in the present application;

[0060] Construction of a standard three-dimensional model and preliminary defect determination:

[0061] A standard three-dimensional model of the part is preset according to the design drawing or standard specification of the part; the standard three-dimensional model defines the size, shape and factory range information of each part of the part;

[0062] The standard three-dimensional model is compared and analyzed with the three-dimensional model of the actual part, the deviation between each point of the standard three-dimensional model and the three-dimensional model of the actual part is calculated, and it is determined whether there is a convex or concave area; the area A and depth H of the convex or concave area are identified, and when the area or depth of the convex or concave area exceeds the preset tolerance range, the area is marked as a potential defect area; the area A and depth H of the convex or concave area are weighted to obtain a defect score S;

[0063] The three-dimensional model of the actual part is divided into several regions according to the structural characteristics and functional importance of the part (for example, independent regions are divided for key stress parts, connection parts, etc.), and a region importance weight wr is set for each region, with a value range of (0, 1], and r represents the number of the region;

[0064] The defect score Sj of each potential defect region in each region and the total number of defect regions M are identified;

[0065] The total number of potential defect regions in the region and the total number of defect regions are weighted to obtain the region defect value D, which is expressed by the formula where j represents the number of the potential defect region, M represents the total number of potential regions, wj represents the severity weight of the potential defect region j, and wm represents the defect quantity weight;

[0066] The region defect values of all divided regions on the part and their corresponding region importance weights are weighted to obtain the comprehensive defect score of the part , which is expressed by the formula where R represents the total number of divided regions, and Dr represents the region defect value corresponding to the region r;

[0067] If the comprehensive defect value is greater than the preset defect threshold value, the part is unqualified; otherwise, the part is qualified;

[0068] Generating a detection report: the basic information of the part (such as the part name, model, batch number, and manufacturer), the qualified rate in the preset production cycle, the qualified trend change, the detailed information of all potential defect regions (including position, area, depth, and defect score), the defect situation of each region (region defect value, region importance weight), and the overall comprehensive defect score are marked as a detection report; the detection report is presented in a combination of visual charts (such as a column chart to show the comparison of region defect values, and a line chart to present the qualified trend change) and written instructions.

[0069] In the present application, the data fusion and processing module further comprises a marking display module;

[0070] The marking display module is used for matching each defect score of the parts with a preset defect marking value range group to obtain a corresponding defect marking when the parts are unqualified; the defect marking is visually marked and displayed on a preset three-dimensional model of the parts; it is to be noted that the preset defect marking value range group includes defect marking value ranges of different levels, for example, slight defects, moderate defects and serious defects correspond to different score intervals; each defect score is compared with the defect marking value range group to determine the category thereof, for example, a certain defect score falls in a serious defect value interval, and a corresponding serious defect marking is matched to provide a basis for subsequent visual marking;

[0071] By using three-dimensional modeling and graphic rendering technology, different levels of defects are marked with different colors, shapes or patterns at the defect positions of the three-dimensional model; for example, serious defects are marked with a prominent red triangle, moderate defects are marked with a yellow circle and slight defects are marked with a blue square, so that the staff can intuitively understand the defect positions and severity and quickly evaluate the quality status of the parts;

[0072] By the marking display module, the abstract defect score can be quickly converted into intuitive visual marking when the parts are unqualified, so that the operator, quality management personnel and engineer can quickly identify the key problem area, the defect status can be clearly understood without complex data analysis, the decision efficiency is significantly improved and misjudgment and missed judgment are reduced.

[0073] The working principle of the large part appearance visual detection equipment provided by the application is as follows:

[0074] Preparation stage: connect the equipment electrical pipeline on the mechanical hand detection mechanism 4, connect the power supply and air pump; place the large part on the support frame 6 of the placing box 2 to complete the preparation work before detection;

[0075] Detection stage: start the mechanical hand detection mechanism 4, the mechanical gripper 7 drives the detection assembly to move to a suitable position; the visual detector at the bottom end of the detection main column 9 starts to preliminarily detect the large part to obtain overall appearance information; at the same time, the laser emission and receiving module starts to work to emit a laser beam to the surface of the part; the laser measurement control unit calculates the distance data of the surface of the part in real time according to the reflected laser signal and generates a three-dimensional model of the actual part; the sliding detection column 11 in the movable sleeve can adjust the extension length and angle according to the detection requirement under the cooperation of the sliding gear 13 and the sliding gear slot 12 driven by the micro motor, so as to further improve the detection precision; the rotating motor 10 drives the transmission disc to rotate, the movable sleeve is rotated through the cooperation of the convex block and the fan-shaped slot, the position of the detection probe is adjusted, the detection of different parts of the part is realized, the complex structure on the part can also be flexibly avoided, and the detection efficiency is improved;

[0076] Data processing and analysis stage: the laser measurement control unit transmits the acquired laser measurement data to the data fusion and processing module, and fuses the data collected by the visual detector; the module analyzes and processes the fused data through a specific algorithm, judges whether the overall and each part of the parts is qualified, and generates a detection report.

[0077] End stage: after the detection is completed, the mechanical hand detection mechanism 4 returns to the original position, the power supply and the air pump are turned off, and the whole detection process is completed.

[0078] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can make equivalent replacement or change according to the technical scheme and the inventive concept of the present application within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A large component appearance visual inspection apparatus comprising a base plate (1), characterized in that, The top surface of the bottom plate (1) is provided with a placing box (2), the side end of the placing box (2) is provided with a sliding guide seat (3), the sliding guide seat (3) is vertically arranged on the bottom plate (1), a mechanical hand detection mechanism (4) is slidably arranged on the sliding guide seat (3), the mechanical hand detection mechanism (4) is composed of a mechanical gripper (7) and a detection assembly, the mechanical gripper (7) is slidably sleeved on the sliding guide seat (3) on one side, the other side of the mechanical gripper (7) is provided with a butt joint seat (8), the detection assembly comprises a detection main column (9), a plurality of movable rotating sleeves, a rotating motor (10) and a visual detector; the visual detector is used for collecting visual detection data of parts; It also includes a laser measurement detection assembly, which includes a laser emission and reception module, a laser measurement control unit, a data fusion and processing module; The laser emission and reception module is used for sending laser signals and receiving laser signals reflected from the surface of the parts; The laser measurement control unit is used for adjusting the laser emission parameters according to the detection requirements; processing and analyzing the received laser signals to obtain the distance information of the part surface and the three-dimensional model of the actual part; wherein the laser emission parameters include the frequency, power and emission angle of the laser emission; The data fusion and processing module is used for fusing and processing the visual detection data and the laser measurement data, analyzing and processing the fused data through a specific algorithm, judging whether the overall part and each part are qualified, and generating a detection report; The detection assembly is composed of a detection main column (9) and a plurality of movable rotating sleeves, the movable rotating sleeves are equidistantly clamped on the detection main column (9), the detection main column (9) is equidistantly provided with a rotating motor (10), the motor shaft of the rotating motor (10) is provided with a transmission disc, the transmission disc is equidistantly provided with a protrusion, the detection main column (9) is provided with a fan-shaped notch for rotating the protrusion, the protrusion of the transmission disc is fixedly connected with the movable rotating sleeve, and the bottom end of the detection main column (9) is provided with a visual detector; The movable rotating sleeve is slidably provided with a sliding detection column (11), the bottom end of the sliding detection column (11) is provided with a laser emission and reception module, the upper end of the sliding detection column (11) is provided with two sliding gears (13) driven by micro motors, the movable rotating sleeve is provided with a sliding gear slot (12) for matching the sliding gear (13), and the movable rotating sleeve is provided with a limiting sliding groove (14) for clamping the sliding detection column (11); Each group of laser emission and reception modules is installed at the position of each movable rotating sleeve corresponding sliding detection column (11), the laser emission device emits a laser beam, which irradiates on the surface of the large part, and the laser receiving device receives the laser signal reflected from the surface of the part; The rotating motor (10) drives the transmission disc to rotate, the movable rotating sleeve is rotated through the cooperation of the protrusion and the fan-shaped notch, the position of the laser emission and reception module is adjusted, the detection of different parts of the part is realized, and the complex structure on the part is avoided.

2. The apparatus according to claim 1, wherein The bottom end of the support column of the placing box (2) is provided with a protective support pad (5), and the upper end of the placing box (2) is provided with a support frame (6) for supporting large parts.

3. The apparatus according to claim 1, wherein The received laser signal is processed and analyzed, specifically: Identify the detection requirements of the parts, including the material, surface characteristics and detection accuracy requirements of the large parts; determine the frequency, power and emission angle of the laser emission according to the detection requirements; Obtain the length of time from laser emission to reception, calculate the distance from the laser emission point to the part surface measurement point using a preset formula according to the propagation speed of laser in air, and record the distance as the measurement distance; An actual three-dimensional model of the part is constructed by the principle of triangulation combined with the emission angle and the relative position relationship of the measurement points.

4. The apparatus according to claim 1, wherein The execution steps of the data fusion and processing module are as follows: Data alignment: using a feature point-based matching algorithm to process the laser measurement data and visual inspection data, extracting feature points from the visual images in the visual inspection data and the laser measurement data, and matching using the feature point descriptors to determine the correspondence between the two, achieving spatial alignment of the data; Constructing a standard three-dimensional model and judging defects: According to the design drawings or standard specifications of the parts, a standard three-dimensional model of the parts is preset; the standard three-dimensional model defines the size, shape and factory range information of each part of the parts; Compare and analyze the standard three-dimensional model with the three-dimensional model of the actual part, calculate the deviation between each point of the standard three-dimensional model and the three-dimensional model of the actual part, and determine whether there are protruding or recessed areas; Identify the area and depth of the protrusions or recesses in the protruding or recessed areas, and mark the area as a potential defect area when the area or depth exceeds the preset tolerance range; Weighted calculation of the area and depth of the protrusions or recesses in the protruding or recessed areas to obtain a defect score; Divide the three-dimensional model of the actual part into several regions according to the structural characteristics and functional importance of the part, and set a regional importance weight for each region; Identify the defect score and total number of potential defect areas in each region; Weighted calculation of all potential defect areas and total number of defect areas in the region to obtain a regional defect value; Weighted processing of the regional defect values of all divided regions on the part and their corresponding regional importance weights to obtain a comprehensive defect score of the part; If the comprehensive defect value is greater than the preset defect threshold value, the part is unqualified; otherwise, the part is qualified; Finally, generate a detection report: Mark the basic information of the part, the qualification rate in the preset production cycle, the change of qualification trend, the detailed information of all potential defect areas, the defect situation of each region and the overall comprehensive defect score as a detection report; the detection report is presented in a combination of visual charts and textual descriptions.

5. The apparatus for visual inspection of large parts according to claim 1, wherein, The data fusion and processing module further comprises a marking display module; The marking display module is used to match the defect score of each defect on the part with the preset defect mark value range group when the part is unqualified, to obtain the corresponding defect mark. Visualizing marking and displaying the defect on a pre-set three-dimensional model of the component.

Images