CCD (Charge Coupled Device) detector for detecting precision hardware and detection method

Abstract

The invention relates to the technical field of hardware detection, in particular to a CCD detection machine for precision hardware detection and a detection method. Performing intersection analysis on the feature model and a preset model, and if the intersection degree between the feature model and the preset model is greater than a preset intersection degree, calibrating the target hardware as a qualified semi-finished product, and enabling the target hardware to flow into a remaining machining center for continuous machining; if the intersection degree between the feature model and a preset model is not greater than a preset intersection degree, performing defect characteristic analysis on the target hardware; if the target hardware is an unqualified semi-finished product, the target hardware is prevented from flowing into the remaining machining centers for continuous machining; otherwise, the target hardware is made to flow into the remaining machining centers for continuous machining. High-precision and automatic hardware detection and defect identification can be realized, the detection efficiency and the product quality are obviously improved, and the production cost can be effectively reduced.

Description

Technical Field

[0001] The invention relates to the technical field of hardware detection, in particular to a CCD detection machine and a detection method for precision hardware detection. Background Art

[0002] With the rapid development of precision manufacturing, higher and higher requirements are put forward for the dimensional accuracy, surface quality and shape integrity of hardware. At present, the detection technology based on machine vision has become an important development direction in the field of precision hardware detection. Among them, CCD detection machine has been widely used in the detection of precision hardware due to its advantages of high resolution, high sensitivity and non-contact detection. However, most of the existing CCD detection machines and detection methods are limited to the detection of simple features of hardware, such as size, contour, etc., and the recognition ability of complex shapes, fine structures and surface defects is limited, which is difficult to meet the growing demand for quality control of precision hardware. In addition, the existing detection methods usually lack the detailed classification of defect types and the consideration of subsequent processing technology, which makes it impossible to effectively guide the subsequent processing process, affecting production efficiency and product quality. Therefore, it is of great practical significance to develop a CCD detection machine and detection method that can collect multi-angle feature information of precision hardware, build feature models and perform intersection analysis with preset models, and then identify different types of defects, and make judgments based on the characteristics of subsequent processing technology, in order to improve the detection accuracy and automation level of precision hardware, reduce production costs, and ensure product quality. Summary of the invention

[0003] The invention overcomes the shortcomings of the prior art and provides a CCD detection machine and a detection method for precision hardware detection.

[0004] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is: The first aspect of the present invention discloses a method for detecting precision hardware, comprising the following steps: Acquire characteristic image information of a target hardware at several preset angles through a CCD camera module on a CCD inspection machine, construct a characteristic model of the target hardware according to the characteristic image information; and acquire a preset model of the target hardware; Performing an intersection analysis on the feature model and the preset model to obtain the degree of intersection between the feature model and the preset model; if the degree of intersection between the feature model and the preset model is greater than the preset degree of intersection, the target hardware is marked as a qualified semi-finished product, and the target hardware flows into the remaining machining center for further processing; If the degree of intersection between the characteristic model and the preset model is not greater than the preset degree of intersection, defect characteristic analysis is performed on the target hardware to obtain a first-class defect area, a second-class defect area and a normal area of ​​the target hardware; Further judge and analyze the target hardware according to the first-class defect area and the second-class defect area of ​​the target hardware and the process characteristics of the remaining machining center required for processing the target hardware to determine whether the target hardware is an unqualified semi-finished product; If the target hardware is an unqualified semi-finished product, the target hardware is prevented from flowing into the remaining machining centers for further processing; otherwise, the target hardware is allowed to flow into the remaining machining centers for further processing.

[0005] Preferably, the feature model is subjected to an intersection analysis with the preset model to obtain the degree of intersection between the feature model and the preset model, specifically: Preset a grid unit size, and construct a three-dimensional grid coordinate system according to the preset grid unit size, and import the feature model and the preset model into the three-dimensional grid coordinate system; Retrieving reference points in the feature model and the preset model so that the reference points of the feature model and the preset model coincide with each other, thereby completing a reference matching process between the feature model and the preset model; Construct a table in a preset format and traverse each grid cell in the three-dimensional grid coordinate system based on the KD tree algorithm; If a grid cell is an idle cell, fill 0 in the blank position of the table; if a grid cell is only occupied by a feature model, fill 1 in the blank position of the table; If a grid cell is occupied only by a preset model, fill in 2 in the blank position of the table; if a grid cell is occupied by both a feature model and a preset model, fill in 3 in the blank position of the table; After all grid cells in the three-dimensional grid coordinate system are traversed by the KD tree algorithm, the total number of cells recorded as 1 in the table is counted; the total number of cells recorded as 2 in the table is counted; the total number of cells recorded as 3 in the table is counted; Divide the total number of values ​​recorded as 3 in the table by the total number of values ​​recorded as 1 in the table, or divide the total number of values ​​recorded as 3 in the table by the total number of values ​​recorded as 2 in the table; and obtain the degree of intersection between the feature model and the preset model.

[0006] Preferably, if the degree of intersection between the characteristic model and the preset model is not greater than the preset degree of intersection, defect characteristic analysis is performed on the target hardware to obtain a first-class defect area, a second-class defect area and a normal area of ​​the target hardware, specifically: If a grid cell is occupied by both the feature model and the preset model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a normal area according to the relative coordinate information between the grid cell and the reference point; If a grid cell is only occupied by the feature model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a type of defect area according to the relative coordinate information between the grid cell and the reference point; If a grid cell is only occupied by a preset model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a second-class defect area according to the relative coordinate information between the grid cell and the reference point.

[0007] Preferably, the target hardware is further judged and analyzed based on the first and second defect areas of the target hardware and the process characteristics of the remaining machining center required for processing the target hardware to determine whether the target hardware is an unqualified semi-finished product, specifically: Obtain a manufacturing process flow chart of the target hardware, determine the remaining machining centers required to process the target hardware after the CCD inspection machine inspection process according to the manufacturing process flow chart, and obtain the process characteristics of each remaining machining center; wherein the process characteristics include process parameters, process types, and process positions; If the process type of a certain residual machining center is a grinding and cutting type, the residual machining center is calibrated as a Class I residual machining center; If the process type of a certain residual machining center is the spray repair processing type, the residual machining center shall be calibrated as a Class II residual machining center; The grinding and cutting processing area for each type of remaining machining center to process the target hardware is determined through the process parameters and process positions of each type of remaining machining center, and is defined as a type of processing area; The spray repair processing area where each of the second-class remaining machining centers processes the target hardware is determined through the process processing parameters and process processing positions of each of the second-class remaining machining centers, and is defined as the second-class processing area.

[0008] The following steps are also included: Obtaining a first-class defect area of ​​a target hardware, and obtaining a second-class defect area of ​​a target hardware; Determining whether the type of processed area can completely cover the type of defective area; If the first type of processing area can completely cover the first type of defective area, it means that the first type of defective area of ​​the target hardware can be completely eliminated by the remaining machining centers; if the first type of processing area cannot completely cover the first type of defective area, it means that the first type of defective area of ​​the target hardware cannot be completely eliminated by the remaining machining centers; Determining whether the second type of processing area can completely cover the second type of defect area; If the second type of processing area can completely cover the second type of defect area, it means that the second type of defect area of ​​the target hardware can be completely eliminated through the remaining machining centers; if the second type of processing area cannot completely cover the second type of defect area, it means that the second type of defect area of ​​the target hardware cannot be completely eliminated through the remaining machining centers.

[0009] The following steps are also included: If there is only one type of defective area in the target hardware and the one type of processing area can completely cover the one type of defective area, the target hardware is regarded as a qualified semi-finished product, and the target hardware is allowed to flow into the remaining machining centers for further processing; if there is only one type of defective area in the target hardware but the one type of processing area cannot completely cover the one type of defective area, the target hardware is marked as an unqualified semi-finished product, and the target hardware is prevented from flowing into the remaining machining centers for further processing; If there are only two types of defective areas in the target hardware and the two types of processing areas can completely cover the two types of defective areas, the target hardware is regarded as a qualified semi-finished product, and the target hardware is allowed to flow into the remaining machining centers for further processing; if there are only two types of defective areas in the target hardware but the two types of processing areas cannot completely cover the two types of defective areas, the target hardware is marked as an unqualified semi-finished product, and the target hardware is prevented from flowing into the remaining machining centers for further processing; If both Class I defect areas and Class II defect areas exist in the target hardware, the target hardware will be regarded as a qualified semi-finished product only when the Class I processing area can completely cover the Class I defect area and the Class II processing area can completely cover the Class II defect area, and the target hardware will be allowed to flow into the remaining machining centers for further processing; otherwise, the target hardware will be marked as an unqualified semi-finished product, and the target hardware will be prevented from flowing into the remaining machining centers for further processing.

[0010] The second aspect of the present invention discloses a CCD inspection machine for precision hardware inspection, which is applied to any one of the precision hardware inspection methods described above, comprising: A testing platform, a CCD camera module and a controller arranged on the testing platform; The first conveyor belt is used to transport the semi-finished hardware parts processed by the previous machining center to the inspection table for passing; The second conveyor belt is used to transport the semi-finished hardware parts after inspection to the subsequent machining center or to the scrap center; The image processing module processes the image collected by the CCD camera module, extracts feature information, and generates feature images; The data storage module stores various data generated during the inspection process, including image data, model data, and defect data.

[0011] The present invention solves the technical defects existing in the background technology, and has the following beneficial effects: This CCD detection machine and detection method for precision hardware detection can realize high-precision, automated hardware detection and defect identification by integrating multiple modules, significantly improve detection efficiency and product quality, and optimize subsequent processing procedures at the same time, ensuring that unqualified products will not flow into subsequent processes, thereby comprehensively improving production efficiency and product quality, and can effectively reduce production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0013] Figure 1 This is the overall method flow chart of this precision hardware testing method; Figure 2 This is a partial method flow chart of this precision hardware testing method; Figure 3 This is a simplified schematic diagram of the structure of this CCD detection machine. DETAILED DESCRIPTION

[0014] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0015] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0016] The first aspect of the present invention discloses a method for detecting precision hardware. Figure 1 As shown, the following steps are included: S102: acquiring characteristic image information of a target hardware at several preset angles through a CCD camera module on a CCD inspection machine, constructing a characteristic model of the target hardware according to the characteristic image information; and acquiring a preset model of the target hardware; The characteristic model is a model diagram of the actual characteristic state of the target hardware after the previous process steps, and the characteristic model represents the actual size and shape of the target hardware; the preset model is a model diagram of the ideal characteristic state of the target hardware after the previous process steps, and the preset model represents the ideal size and shape of the target hardware; S104: performing an intersection analysis on the feature model and the preset model to obtain the degree of intersection between the feature model and the preset model; if the degree of intersection between the feature model and the preset model is greater than the preset degree of intersection, the target hardware is marked as a qualified semi-finished product, and the target hardware flows into the remaining machining center for further processing; S106: If the degree of intersection between the characteristic model and the preset model is not greater than the preset degree of intersection, defect characteristic analysis is performed on the target hardware to obtain a first-class defect area, a second-class defect area, and a normal area of ​​the target hardware; S108: further judging and analyzing the target hardware according to the first-class defect area and the second-class defect area of ​​the target hardware and the process characteristics of the remaining machining center required for processing the target hardware, so as to determine whether the target hardware is an unqualified semi-finished product; S110: If the target hardware component is an unqualified semi-finished product, the target hardware component is prevented from flowing into the remaining machining centers for further processing; otherwise, the target hardware component is allowed to flow into the remaining machining centers for further processing.

[0017] It should be noted that the CCD camera module on the CCD inspection machine captures the image information of the hardware from multiple angles. This image information is used to construct the actual feature model of the hardware. This process ensures that the appearance and shape of the hardware from different viewing angles can be fully and accurately recorded. The "feature model" represents the actual state of the hardware after processing, while the "preset model" is the ideal state that the hardware should achieve. By comparing the two models, it is possible to accurately evaluate whether the hardware meets the design requirements. This comparison is not limited to appearance, but also includes key parameters such as size and shape. By calculating the "degree of intersection" between the feature model and the preset model, the difference between the two can be quantified. If this value exceeds the preset value, it means that the hardware has reached the expected standard and can continue with subsequent processing.

[0018] Otherwise, further analysis and processing are required. Specifically, for products that do not meet the standards, the system will automatically identify their defective areas and divide them into Class I defective areas and Class II defective areas. Based on the specific defects of the hardware and the requirements of the required processing technology, the system will decide whether to allow the product to enter the next stage of processing. This step ensures that only qualified products will be sent to the next process, reducing unnecessary processing procedures for semi-finished waste products and effectively reducing processing costs.

[0019] In summary, this method can significantly improve the quality control level of precision hardware, reduce the defective rate, and improve production efficiency; through automation and intelligent means, it can discover and deal with problems in the semi-finished product stage, promptly scrap semi-finished product waste, reduce unnecessary processing procedures for semi-finished product waste, and effectively reduce processing costs.

[0020] Preferably, the feature model is subjected to an intersection analysis with the preset model to obtain the degree of intersection between the feature model and the preset model, specifically: Preset a grid unit size, and construct a three-dimensional grid coordinate system according to the preset grid unit size, and import the feature model and the preset model into the three-dimensional grid coordinate system; Retrieving reference points in the feature model and the preset model so that the reference points of the feature model and the preset model coincide with each other, thereby completing a reference matching process between the feature model and the preset model; Construct a table in a preset format and traverse each grid cell in the three-dimensional grid coordinate system based on the KD tree algorithm; If a grid cell is an idle cell, fill 0 in the blank position of the table; if a grid cell is only occupied by a feature model, fill 1 in the blank position of the table; If a grid cell is occupied only by a preset model, fill in 2 in the blank position of the table; if a grid cell is occupied by both a feature model and a preset model, fill in 3 in the blank position of the table; After all grid cells in the three-dimensional grid coordinate system are traversed by the KD tree algorithm, the total number of cells recorded as 1 in the table is counted; the total number of cells recorded as 2 in the table is counted; the total number of cells recorded as 3 in the table is counted; Divide the total number of values ​​recorded as 3 in the table by the total number of values ​​recorded as 1 in the table, or divide the total number of values ​​recorded as 3 in the table by the total number of values ​​recorded as 2 in the table; and obtain the degree of intersection between the feature model and the preset model.

[0021] It should be noted that by establishing a three-dimensional grid coordinate system and using the KD tree algorithm for efficient traversal, the intersection degree between the two models is finally calculated. Specifically, first, a preset grid unit size is defined by relevant technicians according to the detection accuracy requirements (such as a cubic grid with a length, width and height of 1 mm), and a three-dimensional grid coordinate system is constructed on this basis. This gridding method can simplify complex three-dimensional models into a series of discrete grid units. After the feature model and the preset model are imported into the three-dimensional grid coordinate system, their reference points need to be found and these reference points overlap. The purpose of this is to ensure that the two models are compared under the same reference frame. The KD tree algorithm is used to traverse all grid cells in the three-dimensional grid coordinate system. For each grid cell, different values ​​are filled in according to whether it is occupied by the feature model and / or the preset model: if the grid cell is an idle cell, fill in 0. If the grid cell is only occupied by the feature model, fill in 1. If the grid cell is only occupied by the preset model, fill in 2. If the grid cell is occupied by both the feature model and the preset model, fill in 3. By counting the total number of 1, 2, and 3 recorded in the table, the degree of overlap between the feature model and the preset model can be obtained. Specifically, there are two ways of calculation: divide the total number of 3 recorded in the table by the total number of 1 recorded in the table. Or, divide the total number of 3 recorded in the table by the total number of 2 recorded in the table. Both calculation methods can be used to represent the degree of intersection between the feature model and the preset model. By accurately mapping the feature model and the preset model to the same three-dimensional grid coordinate system and comparing them grid by grid, this method can achieve very fine model comparison, thereby improving detection accuracy. By filling in different values ​​(0, 1, 2, 3), you can intuitively see which parts are unique to the feature model, which parts are unique to the preset model, and which parts are common to both. This distinction helps to quickly locate and identify defective areas. It not only improves detection accuracy and efficiency, but also provides a clear defect identification mechanism, thereby optimizing the quality control process of precision hardware.

[0022] Preferably, if the degree of intersection between the characteristic model and the preset model is not greater than the preset degree of intersection, defect characteristic analysis is performed on the target hardware to obtain a first-class defect area, a second-class defect area and a normal area of ​​the target hardware, specifically: If a grid cell is occupied by both the feature model and the preset model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a normal area according to the relative coordinate information between the grid cell and the reference point; If a grid cell is only occupied by the feature model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a type of defect area according to the relative coordinate information between the grid cell and the reference point; If a grid cell is only occupied by a preset model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a second-class defect area according to the relative coordinate information between the grid cell and the reference point; Among them, the first type of defect area is the convex hull defect area, and the defects in this area are convex defects, such as bulges, burrs, etc.; the second type of defect area is the concave defect area, and the defects in this area are concave defects, such as cracks, scratches, and pits.

[0023] It should be noted that, according to the previously constructed three-dimensional grid coordinate system, each grid unit is analyzed: if a grid unit is occupied by both the feature model and the preset model, the hardware position corresponding to the grid unit is marked as a normal area. If a grid unit is only occupied by the feature model, the hardware position corresponding to the grid unit is marked as a type I defect area (convex hull defect area). If a grid unit is only occupied by the preset model, the hardware position corresponding to the grid unit is marked as a type II defect area (concave defect area). For each grid unit, the relative coordinate information between it and the reference point is obtained. These relative coordinate information is used to determine the specific location and form of the hardware defect. Through the above-mentioned analysis of the grid unit and the acquisition of relative coordinate information, the different position areas of the hardware can be clearly classified into normal areas, type I defect areas and type II defect areas. Through detailed grid unit analysis and relative coordinate information acquisition, the defect type and location of the hardware can be accurately identified, and very detailed defect information can be provided, which is helpful for subsequent processing. The defect areas are clearly divided into normal areas, type I defect areas and type II defect areas, making the defect classification clearer and more concise. This classification helps to quickly identify and handle different types of defects.

[0024] Preferably, the target hardware is further judged and analyzed based on the first and second defect areas of the target hardware and the process characteristics of the remaining machining center required for processing the target hardware to determine whether the target hardware is an unqualified semi-finished product, such as Figure 2 As shown, specifically: S202: Obtain a manufacturing process flow chart of the target hardware, determine the remaining machining centers required for machining the target hardware after the CCD inspection machine inspection process according to the manufacturing process flow chart, and obtain the process characteristics of each remaining machining center; wherein the process characteristics include process parameters, process types, and process positions; S204: If the process type of a certain remaining machining center is a grinding and cutting process type, the remaining machining center is calibrated as a type I remaining machining center; Among them, the first type of remaining machining centers are machining center workstations that are involved in the subsequent processing of the target hardware and eliminate the materials of the hardware itself, such as edge grinding machine machining centers, polishing machine machining centers, punching machine machining centers, etc.; S206: If the process type of a certain remaining machining center is a spray repair process type, the remaining machining center is calibrated as a second-class remaining machining center; Among them, the second category of remaining machining centers are machining center workstations for filling materials involved in the subsequent processing of target hardware, such as surface painting machining centers, laser cladding machining centers, welding repair machining centers, adhesive filling machining centers, etc.; S208: Determine the grinding and cutting processing area for each of the first-class remaining machining centers to process the target hardware by using the process processing parameters and process processing positions of each of the first-class remaining machining centers, and define it as the first-class processing area; S210: Determine the spray repair processing area where each of the second-class remaining machining centers processes the target hardware through the process processing parameters and process processing positions of each of the second-class remaining machining centers, and define it as the second-class machining area.

[0025] It should be noted that the manufacturing process flow chart of the target hardware is obtained to determine the remaining machining centers that need to be performed after the CCD inspection machine inspection process, and to obtain the process characteristics of these machining centers. According to these characteristics, the remaining machining centers are divided into two categories: the first category of remaining machining centers, which are mainly used for grinding and cutting processing, such as edge grinding machine processing centers, polishing machine processing centers, punching machine processing centers, etc. This type of processing center involves the processing of the hardware's own materials. The second category of remaining machining centers, which are mainly used for spray repair processing, such as surface painting machine processing centers, laser cladding machine processing centers, welding repair machine processing centers, adhesive filling machine processing centers, etc. This type of processing center involves filling the surface defects of hardware.

[0026] By combining the first and second defect areas of the hardware and the process characteristics of the remaining machining centers, the defect types and distribution of the hardware can be analyzed more carefully, which helps to more accurately determine whether the hardware is an unqualified semi-finished product. The remaining machining centers are divided into first and second categories, and the functions and uses of different types of machining centers are clarified, which helps to formulate subsequent processing strategies and ensure that each step of processing can effectively handle specific defect types. At the same time, combining the defect areas of the hardware and the process characteristics of the remaining machining centers, it is possible to more accurately determine whether the hardware is an unqualified semi-finished product. For example, if the first defect area is large and cannot be solved by grinding and cutting processing of the first type of remaining machining centers, then the hardware can be determined to be an unqualified semi-finished product.

[0027] In summary, by analyzing the defective areas of hardware and the process characteristics of the remaining machining centers in detail, refined quality control of hardware is achieved, and the accuracy of identifying unqualified semi-finished products and the rationality of subsequent processing are improved.

[0028] The following steps are also included: Obtaining a first-class defect area of ​​a target hardware, and obtaining a second-class defect area of ​​a target hardware; Determining whether the type of processed area can completely cover the type of defective area; If the first type of processing area can completely cover the first type of defective area, it means that the first type of defective area of ​​the target hardware can be completely eliminated by the remaining machining centers; if the first type of processing area cannot completely cover the first type of defective area, it means that the first type of defective area of ​​the target hardware cannot be completely eliminated by the remaining machining centers; Determining whether the second type of processing area can completely cover the second type of defect area; If the second type of processing area can completely cover the second type of defect area, it means that the second type of defect area of ​​the target hardware can be completely eliminated through the remaining machining centers; if the second type of processing area cannot completely cover the second type of defect area, it means that the second type of defect area of ​​the target hardware cannot be completely eliminated through the remaining machining centers.

[0029] It should be noted that, first of all, the first type of defect area and the second type of defect area of ​​the hardware have been obtained through the previous steps. These defect areas correspond to convex defects (such as bumps, burrs, etc.) and concave defects (such as cracks, scratches, and pits). The first type of processing area is an area specially used for grinding and cutting processing, such as edge grinding machine processing centers, polishing machine processing centers, punching machine processing centers, etc. Determine whether the first type of processing area can completely cover the first type of defect area. If the first type of processing area can completely cover the first type of defect area, it means that these defects can be completely eliminated through the remaining process steps; conversely, if the first type of processing area cannot completely cover the first type of defect area, it means that the defects of the semi-finished hardware products cannot be completely eliminated through subsequent process steps.

[0030] The second-class processing area is an area specially used for spray repair processing, such as surface painting machine processing center, laser cladding machine processing center, welding machine processing center, adhesive filling machine processing center, etc. Determine whether the second-class processing area can completely cover the second-class defect area. If the second-class processing area can completely cover the second-class defect area, it means that these defects can be completely eliminated through the remaining process steps; conversely, if the second-class processing area cannot completely cover the second-class defect area, it means that the defects of the semi-finished hardware products cannot be completely eliminated through subsequent process steps.

[0031] It can be seen that by obtaining clear Class I defect areas and Class II defect areas, the specific defect type and location of the hardware can be accurately determined, which helps to formulate subsequent processing strategies and ensure that each step of processing can effectively handle specific defect types. By judging whether the Class I processing area and the Class II processing area can completely cover the corresponding defect area, the capacity range of the remaining machining center can be clarified, so as to determine whether the subsequent process steps can eliminate the defects in the semi-finished hardware products. If the Class I processing area cannot completely cover the Class I defect area, or the Class II processing area cannot completely cover the Class II defect area, unqualified semi-finished products can be identified in advance, which helps to take timely measures to prevent unqualified products from flowing into subsequent processes, thereby improving overall product quality.

[0032] In summary, this method achieves refined management of hardware defects through accurate defect area acquisition and targeted judgment of processing areas, improves the accuracy of identifying unqualified semi-finished products and the rationality of subsequent processing.

[0033] The following steps are also included: If there is only one type of defective area in the target hardware and the one type of processing area can completely cover the one type of defective area, the target hardware is regarded as a qualified semi-finished product, and the target hardware is allowed to flow into the remaining machining centers for further processing; if there is only one type of defective area in the target hardware but the one type of processing area cannot completely cover the one type of defective area, the target hardware is marked as an unqualified semi-finished product, and the target hardware is prevented from flowing into the remaining machining centers for further processing; If there are only two types of defective areas in the target hardware and the two types of processing areas can completely cover the two types of defective areas, the target hardware is regarded as a qualified semi-finished product, and the target hardware is allowed to flow into the remaining machining centers for further processing; if there are only two types of defective areas in the target hardware but the two types of processing areas cannot completely cover the two types of defective areas, the target hardware is marked as an unqualified semi-finished product, and the target hardware is prevented from flowing into the remaining machining centers for further processing; If both Class I defect areas and Class II defect areas exist in the target hardware, the target hardware will be regarded as a qualified semi-finished product only when the Class I processing area can completely cover the Class I defect area and the Class II processing area can completely cover the Class II defect area, and the target hardware will be allowed to flow into the remaining machining centers for further processing; otherwise, the target hardware will be marked as an unqualified semi-finished product, and the target hardware will be prevented from flowing into the remaining machining centers for further processing.

[0034] It should be noted that whether the hardware is a qualified semi-finished product is judged based on the defective area of ​​the hardware and the processing capacity of the remaining machining center, and whether to continue with subsequent processing is decided.

[0035] Determination of defect areas: Class I defect areas include convex defects (such as bulges, burrs, etc.); Class II defect areas include concave defects (such as cracks, scratches, pits, etc.); Determination of a type of processing area: One type of processing area is an area dedicated to grinding and cutting processing, such as edge grinding machine processing center, polishing machine processing center, punching machine processing center, etc.; Determine whether a type of processing area can completely cover a type of defect area.

[0036] Determination of the second type of processing area: The second-class processing area is an area dedicated to spray repair processing, such as surface painting machine processing center, laser cladding machine processing center, welding machine processing center, adhesive filling machine processing center, etc.; Determine whether the second type of processing area can completely cover the second type of defect area.

[0037] Processing logic in different scenarios: (1) There is only one type of defect area: One type of processing area can be fully covered: the target hardware is regarded as a qualified semi-finished product and flows into the remaining machining centers for further processing; The first type of processing area cannot be fully covered: the target hardware is marked as an unqualified semi-finished product and is prevented from flowing into the remaining machining centers for further processing; (2) There are only two types of defect areas: The second-class processing area can be fully covered: the target hardware is regarded as a qualified semi-finished product and flows into the remaining machining centers for further processing; The second-class processing area cannot be fully covered: the target hardware is marked as an unqualified semi-finished product and is prevented from flowing into the remaining machining centers for further processing; (3) There are both type I defect areas and type II defect areas: If the first-class processing area can completely cover the first-class defect area and the second-class processing area can completely cover the second-class defect area, the target hardware will be regarded as a qualified semi-finished product and flow into the remaining machining centers for further processing; otherwise, the target hardware will be marked as an unqualified semi-finished product and prevented from flowing into the remaining machining centers for further processing.

[0038] It can be seen that through clear defect area judgment and processing area coverage judgment, different defect types of hardware can be accurately handled, which helps to improve the accuracy and efficiency of defect handling. Through the above logical judgment, unqualified semi-finished products can be identified in advance to prevent unqualified products from flowing into subsequent processes, thereby improving the overall product quality. For unqualified semi-finished products, processing can be stopped in time to avoid wasting resources and time and improve production efficiency.

[0039] This method achieves refined quality control of hardware parts through accurate defect area judgment and coverage judgment of processing areas, improves the accuracy of identifying unqualified semi-finished products and the rationality of subsequent processing.

[0040] In addition, constructing a feature model of the target hardware according to the feature image information specifically includes the following steps: Constructing a non-negative matrix decomposition model, importing the characteristic image information into the non-negative matrix decomposition model, and obtaining a basis matrix and a coefficient matrix corresponding to the characteristic image information; Select the first two columns of the basis matrix as basis vectors on the two-dimensional plane, and use each row of the coefficient matrix as a coordinate point on the two-dimensional plane to form a two-dimensional coordinate system; Importing the base matrix and the coefficient matrix into the two-dimensional coordinate system, obtaining eigenvalues ​​of the base matrix and the coefficient matrix, and using the eigenvalues ​​of the base matrix and the coefficient matrix as a set of feature point coordinates; Obtaining a limit coordinate point set of the feature point coordinate set, importing the limit coordinate point set into a Cartesian coordinate system of a CCD camera module for reorganization, and obtaining a partial contour image of the target hardware; Repeat the above steps to perform feature extraction processing on the feature image information of each preset angle to obtain a one-sided contour image of the target hardware at each angle; The partial contour images of the target hardware at various angles are combined and processed to generate a feature model of the target hardware.

[0041] It should be noted that the non-negative matrix factorization (NMF) model is used to import the feature image information into the model to obtain the basis matrix and the coefficient matrix. The first two columns of the basis matrix are selected as the basis vectors on the two-dimensional plane. Each row of the coefficient matrix is ​​used as a coordinate point on the two-dimensional plane to form a two-dimensional coordinate system. The basis matrix and the coefficient matrix are imported into the two-dimensional coordinate system to obtain the eigenvalues ​​of the basis matrix and the coefficient matrix, and these eigenvalues ​​are used as the feature point coordinate collection. The limit coordinate point set of the feature point coordinate collection is obtained, and these limit coordinate point sets are imported into the Cartesian coordinate system of the CCD camera module for reorganization to obtain the one-sided contour map of the target hardware. The feature image information of each preset angle is subjected to feature extraction processing to obtain the one-sided contour map of the target hardware at each angle. The one-sided contour maps of the target hardware at each angle are combined and processed to generate a complete feature model of the target hardware.

[0042] Non-negative matrix factorization (NMF) is used to reduce the dimensionality of high-dimensional feature image information, extract key features, and improve the efficiency and accuracy of feature extraction. Mapping the basis matrix and coefficient matrix into a two-dimensional coordinate system facilitates intuitive observation and understanding of the features of hardware and simplifies the feature analysis process. Through multi-angle feature extraction and combined processing, the three-dimensional feature model of the hardware can be fully reconstructed to improve the integrity of the feature model. By importing the extreme coordinate point set into the Cartesian coordinate system of the CCD camera module for reorganization, the one-sided contour map of the hardware can be accurately reconstructed to improve the reconstruction accuracy. This method can not only efficiently extract and visualize the features of hardware, but also generate a complete feature model through multi-angle reconstruction, improve the integrity and accuracy of the feature model, and thus improve the detection accuracy.

[0043] In addition, the detection method further comprises the following steps: Obtain a manufacturing process flow chart of the target hardware, and determine a pre-processing center for processing the target hardware before the CCD inspection machine inspection process according to the manufacturing process flow chart; Obtaining defect characteristic data generated when each preceding machining center processes and produces hardware parts through a big data network; wherein the defect characteristic data includes the defect location and defect type of the generated defect; If the degree of intersection between the characteristic model and the preset model is not greater than the preset degree of intersection, the first type defect area and the second type defect area of ​​the target hardware are obtained to obtain defect type data of the target hardware; Based on the Pearson correlation coefficient algorithm, the Pearson correlation coefficient between the defect characteristic data generated when each preceding machining center processes and produces the hardware and the defect type data of the target hardware is calculated to obtain a plurality of Pearson correlation coefficients; Constructing a sorting table, importing a plurality of Pearson correlation coefficients into the sorting table for sorting, and obtaining a maximum Pearson correlation coefficient; Obtaining a predecessor machining center corresponding to the maximum Pearson correlation coefficient, defining it as a suspicious predecessor machining center; obtaining real-time working parameters of the suspicious predecessor machining center within a preset time period; Constructing a real-time working parameter curve graph according to the real-time working parameters of the suspected preceding machining center within a preset time period, and obtaining a preset working parameter curve graph of the suspected preceding machining center within the preset time period; An overlap analysis is performed on the real-time working parameter curve graph and the preset working parameter curve graph. If the overlap between the real-time working parameter curve graph and the preset working parameter curve graph is not greater than the preset overlap, the suspicious preceding machining center is defined as a machining center with drifting working parameters, and the working parameters of the machining center with drifting working parameters are corrected.

[0044] It should be noted that the manufacturing process flow chart of the target hardware is obtained, and all the previous machining centers before the CCD inspection machine inspection process are determined according to the flow chart. The defect characteristic data generated by these previous machining centers when producing hardware are collected through the big data network. The defect characteristic data includes the location and type of defects. The defect characteristic data actually collected is compared with the preset model. If the degree of intersection (i.e., similarity) between the two is less than the preset value, it is considered that there is a significant difference and further analysis is required. According to the above comparison results, the first type of defect area and the second type of defect area on the hardware are identified, and the specific defect type data is determined. The Pearson correlation coefficient algorithm is used to calculate the correlation between the defect characteristic data of the previous machining center and the defect type data of the hardware, thereby obtaining a series of Pearson correlation coefficients. These Pearson correlation coefficients are sorted to find the largest correlation coefficient, and then the previous machining center (suspicious previous machining center) that is most likely to cause defects is determined.

[0045] Collect the real-time working parameters of the suspicious preceding machining center within a preset time period, and construct a real-time working parameter curve. Compare and analyze the real-time working parameter curve with the preset working parameter curve. If the overlap between the two is found to be lower than the preset value, it is considered that the working parameters of the machining center have drifted. Finally, perform parameter correction processing on the machining center whose working parameters have drifted to restore normal production. Through automated data analysis and model comparison, the source of problems in the production process can be quickly located, improving the efficiency and accuracy of quality control. The process that affects product quality can be further analyzed based on the test results, so that the problematic process can be adjusted, which can effectively reduce the scrap rate.

[0046] The second aspect of the present invention discloses a CCD inspection machine for precision hardware inspection, which is applied to any one of the precision hardware inspection methods described above, such as Figure 3 As shown, including: A testing platform 8011, a CCD camera module 8012 and a controller 8013 arranged on the testing platform; The first conveyor belt 8014 is used to convey the semi-finished hardware parts processed by the previous machining center to the inspection table for passing; The second conveyor belt 8015 is used to convey the semi-finished hardware parts after inspection to the subsequent machining center or to the scrapping center; The image processing module processes the image collected by the CCD camera module, extracts feature information, and generates feature images; The data storage module stores various data generated during the inspection process, including image data, model data, and defect data.

[0047] In summary, this CCD inspection machine and inspection method for precision hardware inspection can achieve high-precision, automated hardware inspection and defect identification by integrating multiple modules, significantly improving inspection efficiency and product quality, while optimizing subsequent processing procedures to ensure that unqualified products do not flow into subsequent processes, thereby comprehensively improving production efficiency and product quality, and effectively reducing production costs.

[0048] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for detecting precision hardware, characterized in that: The following steps are involved: Acquire characteristic image information of a target hardware at several preset angles through a CCD camera module on a CCD inspection machine, construct a characteristic model of the target hardware according to the characteristic image information; and acquire a preset model of the target hardware; Performing an intersection analysis on the feature model and the preset model to obtain the degree of intersection between the feature model and the preset model; if the degree of intersection between the feature model and the preset model is greater than the preset degree of intersection, the target hardware is marked as a qualified semi-finished product, and the target hardware flows into the remaining machining center for further processing; If the degree of intersection between the characteristic model and the preset model is not greater than the preset degree of intersection, defect characteristic analysis is performed on the target hardware to obtain a first-class defect area, a second-class defect area and a normal area of ​​the target hardware; Further judge and analyze the target hardware according to the first-class defect area and the second-class defect area of ​​the target hardware and the process characteristics of the remaining machining center required for processing the target hardware to determine whether the target hardware is an unqualified semi-finished product; If the target hardware is an unqualified semi-finished product, the target hardware is prevented from flowing into the remaining machining centers for further processing; otherwise, the target hardware is allowed to flow into the remaining machining centers for further processing.

2. A method for detecting precision hardware according to claim 1, characterized in that: The feature model and the preset model are subjected to intersection analysis to obtain the degree of intersection between the feature model and the preset model, specifically: Preset a grid unit size, and construct a three-dimensional grid coordinate system according to the preset grid unit size, and import the feature model and the preset model into the three-dimensional grid coordinate system; Retrieving reference points in the feature model and the preset model so that the reference points of the feature model and the preset model coincide with each other, thereby completing a reference matching process between the feature model and the preset model; Construct a table in a preset format and traverse each grid cell in the three-dimensional grid coordinate system based on the KD tree algorithm; If a grid cell is an idle cell, fill 0 in the blank position of the table; if a grid cell is only occupied by a feature model, fill 1 in the blank position of the table; If a grid cell is occupied only by a preset model, fill in 2 in the blank position of the table; If a grid cell is occupied by both a feature model and a preset model, fill in 3 in the blank position of the table; After all grid cells in the three-dimensional grid coordinate system are traversed by the KD tree algorithm, the total number of cells recorded as 1 in the table is counted; Count the total number of 2s recorded in the table; count the total number of 3s recorded in the table; Divide the total number of 3s recorded in the table by the total number of 1s recorded in the table, or divide the total number of 3s recorded in the table by the total number of 2s recorded in the table; The degree of intersection between the feature model and the preset model is obtained.

3. A method for detecting precision hardware according to claim 2, characterized in that: If the degree of intersection between the characteristic model and the preset model is not greater than the preset degree of intersection, the defect characteristic analysis is performed on the target hardware to obtain the first type defect area, the second type defect area and the normal area of ​​the target hardware, specifically: If a grid cell is occupied by both the feature model and the preset model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a normal area according to the relative coordinate information between the grid cell and the reference point; If a grid cell is only occupied by the feature model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a type of defect area according to the relative coordinate information between the grid cell and the reference point; If a grid cell is only occupied by a preset model, the relative coordinate information between the grid cell and the reference point is obtained, and the position area corresponding to the target hardware is marked as a second-class defect area according to the relative coordinate information between the grid cell and the reference point.

4. A method for detecting precision hardware according to claim 1, characterized in that: According to the first and second defect areas of the target hardware and combined with the process characteristics of the remaining machining center required to process the target hardware, the target hardware is further judged and analyzed to determine whether the target hardware is an unqualified semi-finished product, specifically: Obtain a manufacturing process flow chart of the target hardware, determine the remaining machining centers required to process the target hardware after the CCD inspection machine inspection process according to the manufacturing process flow chart, and obtain the process characteristics of each remaining machining center; wherein the process characteristics include process parameters, process types, and process positions; If the process type of a certain residual machining center is a grinding and cutting type, the residual machining center is calibrated as a Class I residual machining center; If the process type of a certain residual machining center is the spray repair processing type, the residual machining center shall be calibrated as a Class II residual machining center; The grinding and cutting processing area for each type of remaining machining center to process the target hardware is determined through the process parameters and process positions of each type of remaining machining center, and is defined as a type of processing area; The spray repair processing area where each of the second-class remaining machining centers processes the target hardware is determined through the process processing parameters and process processing positions of each of the second-class remaining machining centers, and is defined as the second-class processing area.

5. A method for detecting precision hardware according to claim 4, characterized in that: The following steps are also included: Obtaining a first-class defect area of ​​a target hardware, and obtaining a second-class defect area of ​​a target hardware; Determining whether the type of processed area can completely cover the type of defective area; If the first type of processing area can completely cover the first type of defective area, it means that the first type of defective area of ​​the target hardware can be completely eliminated by the remaining machining centers; if the first type of processing area cannot completely cover the first type of defective area, it means that the first type of defective area of ​​the target hardware cannot be completely eliminated by the remaining machining centers; Determining whether the second type of processing area can completely cover the second type of defect area; If the second type of processing area can completely cover the second type of defect area, it means that the second type of defect area of ​​the target hardware can be completely eliminated through the remaining machining centers; if the second type of processing area cannot completely cover the second type of defect area, it means that the second type of defect area of ​​the target hardware cannot be completely eliminated through the remaining machining centers.

6. A method for detecting precision hardware according to claim 5, characterized in that: The following steps are also included: If there is only one type of defective area in the target hardware and the one type of processing area can completely cover the one type of defective area, the target hardware is regarded as a qualified semi-finished product, and the target hardware is allowed to flow into the remaining machining centers for further processing; if there is only one type of defective area in the target hardware but the one type of processing area cannot completely cover the one type of defective area, the target hardware is marked as an unqualified semi-finished product, and the target hardware is prevented from flowing into the remaining machining centers for further processing; If there are only two types of defective areas in the target hardware and the two types of processing areas can completely cover the two types of defective areas, the target hardware is regarded as a qualified semi-finished product, and the target hardware is allowed to flow into the remaining machining centers for further processing; if there are only two types of defective areas in the target hardware but the two types of processing areas cannot completely cover the two types of defective areas, the target hardware is marked as an unqualified semi-finished product, and the target hardware is prevented from flowing into the remaining machining centers for further processing; If both Class I defect areas and Class II defect areas exist in the target hardware, the target hardware will be regarded as a qualified semi-finished product only when the Class I processing area can completely cover the Class I defect area and the Class II processing area can completely cover the Class II defect area, and the target hardware will be allowed to flow into the remaining machining centers for further processing; otherwise, the target hardware will be marked as an unqualified semi-finished product, and the target hardware will be prevented from flowing into the remaining machining centers for further processing.

7. A CCD inspection machine for precision hardware inspection, applied to a precision hardware inspection method according to any one of claims 1 to 6, characterized in that: include: A testing platform, a CCD camera module and a controller arranged on the testing platform; The first conveyor belt is used to transport the semi-finished hardware parts processed by the previous machining center to the inspection table for passing; The second conveyor belt is used to transport the semi-finished hardware parts after inspection to the subsequent machining center or to the scrap center; The image processing module processes the image collected by the CCD camera module, extracts feature information, and generates feature images; The data storage module stores various data generated during the inspection process, including image data, model data, and defect data.

Images