Production line visual processing anti-repetition method

Abstract

The invention discloses an assembly line visual processing anti-duplication method. The method comprises the following steps: acquiring a visual identification image p (n); determining identification data and visual identification coordinates of the identification data according to the visual identification image p (n); comparing the visual identification image p (n) with a previous visual identification image p (n-1), and determining a repeated identification area as an overlapping area; screening and rejecting identification data corresponding to the visual identification coordinates in the overlapping region; setting a moving distance tolerance; and based on the moving distance tolerance, screening residual identification data in the visual identification image p (n) through the previous visual identification image p (n-1) and the (n-2) th visual identification image p (n-2), and removing the identification data meeting the two screening conditions. According to the invention, the problem of repeated processing is avoided on the basis of not passing visual inspection, the product detection efficiency is improved, and meanwhile, the situation that visual inspection cannot be passed in a severe field environment is effectively prevented.

Description

Technical Field

[0001] The present application relates to the field of laser marking technology, and more specifically, to an anti-weighting method for assembly line visual processing. Background Art

[0002] The visual processing anti-duplicate technology of assembly line is a technology to identify and verify products during the manufacturing process, which is mainly detected by visual inspection methods. However, the visual inspection method has high requirements for the technical environment, and the technical solution is complex and has high technical requirements for personnel.

[0003] This technology is commonly used in automated production and processing lines to ensure product consistency and traceability throughout the entire production process. The importance of assembly line visual processing anti-duplicate is becoming increasingly prominent. It can quickly detect abnormal situations and take measures to correct them by real-time monitoring of product and production process data, thereby improving production efficiency and quality. In addition, assembly line visual processing anti-duplicate can also help companies achieve visualization and intelligent management, and improve the flexibility and adaptability of production lines. However, due to the complexity and diversity of production processes, assembly line visual processing anti-duplicate faces many challenges, such as the accuracy and real-time nature of data collection and processing, rapid identification and response to abnormal situations, and protection of personnel and equipment safety. Therefore, the development of efficient assembly line visual processing anti-duplicate technology and solutions is of great significance to enhancing the competitiveness of the manufacturing industry and responding to future challenges.

[0004] Therefore, the prior art has defects and is in urgent need of improvement. Summary of the invention

[0005] In view of the above problems, the purpose of the present invention is to provide a method for preventing weight loss through visual processing of pipelines, which can improve processing efficiency, lower the threshold of visual weight loss prevention, and solve the problem that the visual detection method cannot effectively prevent weight loss due to the harsh on-site environment. At the same time, the method is simple to use and does not require complex debugging such as visual detection.

[0006] The first aspect of the present invention provides a pipeline visual processing anti-weighting method, comprising: Get the visual recognition image p (n) ; According to the visual recognition image p (n) determining identification data and visual identification coordinates of the identification data; The visual recognition image p (n) and the previous visual recognition image p (n-1) Comparison is performed and the repeated identification areas are determined as overlapping areas; Screening and removing the recognition data corresponding to the visual recognition coordinates in the overlapping area; Set the moving distance tolerance; Based on the moving distance tolerance, the previous visual recognition image p is respectively (n-1) and the n-2th visual recognition image p (n-2) For visual recognition image p (n) The remaining identification data are screened and the identification data that meet the two screening conditions are eliminated.

[0007] In this solution, the acquisition of visual recognition image p (n) ,include: Set the preset movement distance based on the length of the camera's field of view; Get the moving distance of the motor shaft; When the motor shaft moves to a preset distance, the visual recognition image p is obtained. (n) .

[0008] This plan also includes: The preset moving distance L satisfies: ; Wherein, W is the length of the camera field of view and D is the length of a single recognition data.

[0009] In this solution, the setting of the moving distance tolerance includes: The moving distance tolerance is calculated by the angle difference a between the camera and the motor shaft; the moving distance tolerance includes a lateral tolerance Rx and a longitudinal tolerance Ry; The calculation method of the lateral tolerance Rx and the longitudinal tolerance Ry is expressed by the formula: ; .

[0010] In this solution, based on the moving distance tolerance, the previous visual recognition image p (n-1) and the n-2th visual recognition image p (n-2) For visual recognition image p (n) The remaining identification data is screened and the identification data that meets the two screening conditions is eliminated, including: According to the previous visual recognition image p (n-1) The first theoretical coordinates of the recognition data are determined by the visual recognition coordinates of the recognition data and the preset moving distance, and the recognition data that meets the screening conditions are marked by analyzing the first theoretical coordinates and the visual recognition coordinates of the same recognition data; According to the n-2th visual recognition image p (n-2) The visual recognition coordinates of the recognition data and the preset moving distance are used to determine the second theoretical coordinates of the recognition data, and the recognition data that meets the screening conditions are marked by analyzing the second theoretical coordinates and the visual recognition coordinates of the same recognition data; Count the annotation times of each recognized data, and eliminate the recognized data with two annotation times.

[0011] In this solution, the first theoretical coordinate of the recognized data is determined according to the visual recognition coordinates and the preset moving distance of the recognized data in the previous visual recognition image p (n-1) By analyzing the first theoretical coordinate and the visual recognition coordinate of the same recognized data, the recognized data that meets the screening conditions is marked, including: Determine the center coordinate of the recognized data in the previous visual recognition image p (n-1) as the first reference coordinate (x0, y0); Determine the first theoretical coordinate (x1, y1) of the recognized data after moving once according to the first reference coordinate (x0, y0) and the preset moving distance; Based on the visual recognition coordinates (x2, y2) of the recognized data in the visual recognition image p (n) analyze the first theoretical coordinate (x1, y1), the visual recognition coordinate (x2, y2) and the first distance tolerance of the same recognized data to determine whether the same recognized data meets the screening conditions; the first distance tolerance is the moving distance tolerance (Rx, Ry); When |x1 - x2| < Rx and |y1 - y2| < Ry are satisfied simultaneously, it is determined that the same recognized data meets the screening conditions.

[0012] In this solution, the second theoretical coordinate of the recognized data is determined according to the visual recognition coordinates and the preset moving distance of the recognized data in the (n - 2)th visual recognition image p (n-2) By analyzing the second theoretical coordinate and the visual recognition coordinate of the same recognized data, the recognized data that meets the screening conditions is marked, including: Determine the center coordinate of the recognized data in the (n - 2)th visual recognition image p (n-2) as the second reference coordinate (x0', y0'); Determine the second theoretical coordinate (x3, y3) of the recognized data after moving twice according to the second reference coordinate (x0', y0') and the preset moving distance; Calculate the second distance tolerance (Rx', Ry') of the recognized data according to the moving distance tolerance; Based on the visual recognition coordinates (x2, y2) of the recognized data in the visual recognition image p (n) analyze the second theoretical coordinate (x3, y3), the visual recognition coordinate (x2, y2) and the second distance tolerance (Rx', Ry') of the same recognized data to determine whether the same recognized data meets the screening conditions; When both |x3 - x2| < Rx and |y3 - y2| < Ry are satisfied, it is determined that the same recognition data meets the screening conditions.

[0013] In this solution, it further includes: The calculation method of the second distance tolerance (Rx', Ry') is expressed by the formula: ; .

[0014] The present invention discloses a method for preventing duplicate processing in pipeline vision processing. The method includes: obtaining a vision recognition image p (n) ; determining recognition data and the vision recognition coordinates of the recognition data according to the vision recognition image p (n) ; comparing the vision recognition image p (n) with the previous vision recognition image p (n-1) , and determining the overlapping area as the duplicate recognition area; screening and eliminating the recognition data corresponding to the vision recognition coordinates in the overlapping area; setting a moving distance tolerance; based on the moving distance tolerance, respectively screening the remaining recognition data in the vision recognition image p (n-1) through the previous vision recognition image p (n-2) and the (n - 2)th vision recognition image p (n) , and eliminating the recognition data that meets the two screening conditions. The present invention avoids the occurrence of duplicate processing problems without visual detection, improves the product detection efficiency, and effectively prevents duplication in the case where the on-site environment is harsh and visual detection cannot be passed. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Shows a flowchart of a method for preventing duplicate processing in pipeline vision processing provided by the present invention; Figure 2 Shows a flowchart of a method for obtaining a vision recognition image provided by the present invention; Figure 3 Shows a flowchart of a method for screening and eliminating the remaining recognition data in the vision recognition image provided by the present invention; Figure 4 Shows a schematic diagram of the (n - 1)th vision recognition image provided by the present invention; Figure 5 Shows a schematic diagram of the nth vision recognition image provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] In order to more clearly understand the above objects, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0017] 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.

[0018] Figure 1 A flow chart of an assembly line visual processing anti-duplicate method provided by the present invention is shown.

[0019] like Figure 1 As shown, the present invention discloses a pipeline visual processing anti-weighting method, comprising: S102, obtaining a visual recognition image p (n) ; S104, based on the visual recognition image p (n) determining identification data and visual identification coordinates of the identification data; S106, visual recognition image p (n) and the previous visual recognition image p (n-1) Comparison is performed and the repeated identification areas are determined as overlapping areas; S108, filtering and removing the recognition data corresponding to the visual recognition coordinates in the overlapping area; S110, setting the moving distance tolerance; S112, based on the moving distance tolerance, respectively through the previous visual recognition image p (n-1) and the n-2th visual recognition image p (n-2) For visual recognition image p (n) The remaining identification data are screened and the identification data that meet the two screening conditions are eliminated.

[0020] According to an embodiment of the present invention, in a scenario where unidirectional motor movement is combined with visual recognition positioning, the conveyor belt is controlled by a motor so that the product on the conveyor belt moves with the conveyor belt, and a visual recognition image is acquired by the moving distance of the motor shaft. When the moving distance of the motor shaft meets the preset moving distance, a visual recognition image is acquired by a detection device such as a camera. The recognition data in the visual recognition image is the recognition image of a single product, and the center coordinates of the recognition image of each product are determined as the visual recognition coordinates of the recognition data corresponding to the product. The current visual recognition image p is calibrated according to the preset moving distance. (n) and the previous visual recognition image p (n-1) To analyze, the current (nth) visual recognition image p (n) The schematic diagram is as follows Figure 5 As shown, the last (n-1th) visual recognition image p (n-1) The schematic diagram is as follows Figure 4 As shown, Figure 4The middle arrow indicates that the conveyor belt moves from left to right. The repeated recognition areas that appear in the two recognition processes are determined as overlapping areas. (n) The recognition data whose visual recognition coordinates are in the overlapping area are screened out to avoid repeated recognition.

[0021] Figure 2 A flow chart of the visual recognition image acquisition method provided by the present invention is shown.

[0022] like Figure 2 As shown, according to an embodiment of the present invention, a visual recognition image p is obtained. (n) ,include: S202, setting a preset moving distance based on the length of the camera's field of view; S204, obtaining the moving distance of the motor shaft; S206, when the motor shaft movement distance meets the preset movement distance, obtain the visual recognition image p (n) .

[0023] It should be noted that the preset moving distance is smaller than the length of the camera's field of view (i.e., the length of the visual recognition image). By setting a reasonable moving distance, it is ensured that after the motor shaft moves, part of the previous visual recognition image appears within the recognition range of the current visual recognition image. Those skilled in the art can adjust the value of the preset moving distance according to actual needs. The smaller the value of the preset moving distance, the higher the accuracy of preventing missed recognition.

[0024] According to an embodiment of the present invention, it also includes: The preset moving distance L satisfies: ; Wherein, W is the length of the camera field of view and D is the length of a single recognition data.

[0025] It should be noted that the system's default preset moving distance is at least greater than half the length of the camera's field of view. To ensure that adjacent visual recognition images acquired have overlapping areas and that there is at least one complete product's identification data in the overlapping area, the preset moving distance also needs to be less than the length of the camera's field of view minus the length of the identification data of a single (product) on the assembly line.

[0026] According to an embodiment of the present invention, setting the moving distance tolerance includes: The moving distance tolerance is calculated by the angle difference a between the camera and the motor shaft; the moving distance tolerance includes the lateral tolerance Rx and the longitudinal tolerance Ry; The calculation method of the lateral tolerance Rx and the longitudinal tolerance Ry is expressed by the formula: ; .

[0027] It should be noted that the moving distance tolerance includes the lateral tolerance Rx in the lateral direction and the longitudinal tolerance Ry in the longitudinal direction. The moving distance tolerance is represented by (Rx, Ry). The moving distance tolerance is calculated by the angle difference a between the camera and the motor shaft in the assembly line in the horizontal direction or the vertical direction.

[0028] Figure 3 A flow chart of a method for screening and eliminating remaining recognition data in a visual recognition image provided by the present invention is shown.

[0029] like Figure 3 As shown, according to an embodiment of the present invention, based on the moving distance tolerance, the previous visual recognition image p is respectively (n-1) and the n-2th visual recognition image p (n-2) For visual recognition image p (n) The remaining identification data is screened and the identification data that meets the two screening conditions is eliminated, including: S302, based on the previous visual recognition image p (n-1) The first theoretical coordinates of the recognition data are determined by the visual recognition coordinates of the recognition data and the preset moving distance, and the recognition data that meets the screening conditions are marked by analyzing the first theoretical coordinates and the visual recognition coordinates of the same recognition data; S304, based on the n-2th visual recognition image p (n-2) The visual recognition coordinates of the recognition data and the preset moving distance are used to determine the second theoretical coordinates of the recognition data, and the recognition data that meets the screening conditions are marked by analyzing the second theoretical coordinates and the visual recognition coordinates of the same recognition data; S306, counting the number of times each piece of recognition data is marked, and eliminating recognition data that has been marked twice.

[0030] It should be noted that, based on the determination of the overlapping area, the previous visual recognition image p (n-1) and the n-2th visual recognition image p (n-2) For the current visual recognition image p (n) The remaining identification data are screened twice, and the identification data that meet the screening conditions in both screenings are eliminated.

[0031] First, the previous visual recognition image p (n-1) and the n-2th visual recognition image p (n-2) The center coordinates of each recognition data in the image are used as reference coordinates, and the number of moves and the preset moving distance are combined to determine the position of each recognition data in the current visual recognition image p. (n)Based on the theoretical coordinate data therein, calculate the coordinate difference between the theoretical coordinate data and the visual recognition coordinates of the same recognition data. Label the recognition data where the horizontal coordinate difference meets the horizontal tolerance and the vertical coordinate difference meets the vertical tolerance, and determine that the recognition data meets the screening conditions. Count the number of labels for each recognition data in the current visual recognition image p (n) in it, and eliminate the recognition data with two labeling times.

[0032] According to an embodiment of the present invention, based on the visual recognition coordinates of the recognition data in the previous visual recognition image p (n-1) and the preset moving distance, determine the first theoretical coordinates of the recognition data. By analyzing the first theoretical coordinates and the visual recognition coordinates of the same recognition data, label the recognition data that meets the screening conditions, including: Determine the center coordinates of the recognition data in the previous visual recognition image p (n-1) as the first reference coordinates (x0, y0); According to the first reference coordinates (x0, y0) and the preset moving distance, determine the first theoretical coordinates (x1, y1) after the recognition data moves once; Based on the visual recognition coordinates (x2, y2) of the recognition data in the visual recognition image p (n) analyze the first theoretical coordinates (x1, y1), the visual recognition coordinates (x2, y2) and the first distance tolerance of the same recognition data to determine whether the same recognition data meets the screening conditions; the first distance tolerance is the moving distance tolerance (Rx, Ry); When both |x1 - x2| < Rx and |y1 - y2| < Ry are satisfied, determine that the same recognition data meets the screening conditions.

[0033] It should be noted that when performing the first screening on the recognition data in the visual recognition image p (n) , use the previous visual recognition image p (n-1) as the comparison standard. Determine the center coordinates of the recognition data in the previous visual recognition image p (n-1) as the first reference coordinates (x0, y0), and determine the first theoretical coordinates (x1, y1) after the recognition data moves once according to the preset moving distance L, where x1 = x0 + L and y1 = y0. Calculate the distance differences between the first theoretical coordinates (x1, y1) and the visual recognition coordinates (x2, y2) in the horizontal and vertical directions, represented by (|x1 - x2|, |y1 - y2|), and compare this distance difference (|x1 - x2|, |y1 - y2|) with the first distance tolerance (Rx, Ry). When both |x1 - x2| < Rx and |y1 - y2| < Ry are satisfied, determine that the corresponding recognition data is duplicate recognition and that the recognition data meets the screening conditions for the first screening.

[0034] Among them, due to the current visual recognition image p (n) is obtained by moving the previous visual recognition image p (n-1) once, the first distance tolerance can be represented by the moving distance tolerance, that is, the first distance tolerance is (Rx, Ry).

[0035] ; .

[0036] According to the embodiment of the present invention, the second theoretical coordinates of the recognition data are determined according to the visual recognition coordinates of the recognition data in the (n - 2)-th visual recognition image p (n-2) and the preset moving distance. By analyzing the second theoretical coordinates and the visual recognition coordinates of the same recognition data, the recognition data that meets the screening conditions is marked, including: Determine the center coordinates of the recognition data in the (n - 2)-th visual recognition image p (n-2) as the second reference coordinates (x0', y0'); Determine the second theoretical coordinates (x3, y3) of the recognition data after moving twice according to the second reference coordinates (x0', y0') and the preset moving distance; Calculate the second distance tolerance (Rx', Ry') of the recognition data according to the moving distance tolerance; Based on the visual recognition coordinates (x2, y2) of the recognition data in the visual recognition image p (n) , analyze the second theoretical coordinates (x3, y3), the visual recognition coordinates (x2, y2) and the second distance tolerance (Rx', Ry') of the same recognition data to determine whether the same recognition data meets the screening conditions; When |x3 - x2| < Rx and |y3 - y2| < Ry are satisfied simultaneously, it is determined that the same recognition data meets the screening conditions.

[0037] It should be noted that when the recognition data in the visual recognition image p (n) is screened for the second time, the (n - 2)-th visual recognition image p (n-2) is used as the comparison standard, and the (n - 2)-th visual recognition image p (n-2)The central coordinates of the recognition data are determined as the second reference coordinates (x0', y0'). According to the preset moving distance L, the second theoretical coordinates (x3, y3) after the recognition data moves once are determined, where x3 = x0' + 2L and y3 = y0'. Calculate the distance differences in the horizontal and vertical directions between the second theoretical coordinates (x3, y3) and the visual recognition coordinates (x2, y2), which is represented by (|x3 - x2|, |y3 - y2|). Compare the distance difference (|x3 - x2|, |y3 - y2|) with the second distance tolerance (Rx', Ry'). When both |x1 - x2| < Rx' and |y1 - y2| < Ry' are satisfied, determine that the corresponding recognition data is duplicate recognition, and determine that the recognition data meets the screening conditions for the second screening.

[0038] According to an embodiment of the present invention, it further includes: The calculation method of the second distance tolerance (Rx', Ry') is expressed by the formula: ; .

[0039] It should be noted that the distance tolerance is determined according to the product of the number of movements and the movement distance tolerance. The second distance tolerance is determined by the visual recognition image p of the (n - 2)th time (n-2) and the current visual recognition image p (n) Determined, from the visual recognition image p (n-2) to the visual recognition image p (n) There are two movements. Therefore, Ry' = 2×Ry and Rx' = 2×Rx.

[0040] The information involved in this application (including but not limited to user device information, user personal information, etc.), data (including but not limited to data for analysis, stored data, displayed data, etc.), and signals (including but not limited to signals transmitted between the user terminal and other devices, etc.) are all authorized by the user or fully authorized by all parties. And the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions. For example, the "visual recognition image", "motor shaft movement distance", etc. involved in this disclosure are all obtained under full authorization.

[0041] The present invention discloses a method for preventing duplicate in pipeline visual processing. The method includes: obtaining the visual recognition image p (n) ; determining the recognition data and the visual recognition coordinates of the recognition data according to the visual recognition image p (n) ; the visual recognition image p (n) and the previous visual recognition image p (n-1)Compare and determine the repeated recognition area as the overlapping area; filter and remove the recognition data corresponding to the visual recognition coordinates in the overlapping area; set the moving distance tolerance; based on the moving distance tolerance, respectively use the previous visual recognition image p (n-1) and the n-2th visual recognition image p (n-2) For visual recognition image p (n) The remaining identification data is screened, and the identification data that meets the two screening conditions is eliminated. The present invention avoids the problem of repeated processing without visual inspection, improves product inspection efficiency, and effectively prevents duplication when the on-site environment is harsh and visual inspection is not possible.

[0042] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0043] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0044] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0045] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0046] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

Claims

1. A method for preventing weight loss in pipeline visual processing, characterized in that: Including: Get the visual recognition image p (n) ; According to the visual recognition image p (n) determining identification data and visual identification coordinates of the identification data; The visual recognition image p (n) and the previous visual recognition image p (n-1) Comparison is performed and the repeated identification areas are determined as overlapping areas; Screening and eliminating the recognition data corresponding to the visual recognition coordinates within the overlapping area; Setting the movement distance tolerance; Based on the moving distance tolerance, the previous visual recognition image p is respectively (n-1) and the n-2th visual recognition image p (n-2) For visual recognition image p (n) The remaining identification data are screened and the identification data that meet the two screening conditions are eliminated.

2. The method for preventing weight loss in pipeline visual processing according to claim 1, characterized in that: The obtaining of the visual recognition image p (n) ,include: Setting a preset movement distance based on the length of the camera's field of view; Obtaining the movement distance of the motor shaft; When the motor shaft moves to a preset distance, the visual recognition image p is obtained. (n) .

3. The method for preventing weight loss in pipeline visual processing according to claim 2, characterized in that: Also including: The preset movement distance L satisfies: ； Where W is the length of the camera's field of view and D is the length of a single recognition data.

4. The method for preventing weight loss in pipeline visual processing according to claim 1, characterized in that: The setting of the movement distance tolerance includes: Calculating the movement distance tolerance through the angle difference a between the camera and the motor shaft; the movement distance tolerance includes a horizontal tolerance Rx and a vertical tolerance Ry; The calculation methods of the horizontal tolerance Rx and the vertical tolerance Ry are expressed by the formula: ； 。 5. The method for preventing weight loss in pipeline visual processing according to claim 1, characterized in that: Based on the moving distance tolerance, the last visual recognition image p is respectively (n-1) and the n-2th visual recognition image p (n-2) For visual recognition image p (n) The remaining identification data is screened and the identification data that meets the two screening conditions is eliminated, including: According to the previous visual recognition image p (n-1) The first theoretical coordinates of the recognition data are determined by the visual recognition coordinates of the recognition data and the preset moving distance, and the recognition data that meets the screening conditions are marked by analyzing the first theoretical coordinates and the visual recognition coordinates of the same recognition data; According to the n-2th visual recognition image p (n-2) The visual recognition coordinates of the recognition data and the preset moving distance are used to determine the second theoretical coordinates of the recognition data, and the recognition data that meets the screening conditions are marked by analyzing the second theoretical coordinates and the visual recognition coordinates of the same recognition data; Counting the annotation times of each recognition data and eliminating the recognition data with two annotation times.

6. The method for preventing weight loss in pipeline visual processing according to claim 5, characterized in that: The method according to the previous visual recognition image p (n-1) The first theoretical coordinates of the recognition data are determined by the visual recognition coordinates of the recognition data and the preset moving distance, and the recognition data that meets the screening conditions are marked by analyzing the first theoretical coordinates and the visual recognition coordinates of the same recognition data, including: The previous visual recognition image p (n-1) The center coordinates of the identified data are determined as the first reference coordinates (x0, y0); Determining the first theoretical coordinates (x1, y1) after the recognition data moves once according to the first reference coordinates (x0, y0) and the preset movement distance; Based on visual recognition image p (n) The visual recognition coordinates (x2, y2) of the recognition data are analyzed to determine whether the same recognition data meets the screening condition; the first distance tolerance is the moving distance tolerance (Rx, Ry); When both |x1 - x2| < Rx and |y1 - y2| < Ry are satisfied, it is determined that the same recognition data meets the screening conditions.

7. The method for preventing weight loss in pipeline visual processing according to claim 5, characterized in that: The n-2th visual recognition image p (n-2) The visual recognition coordinates of the recognition data and the preset moving distance are used to determine the second theoretical coordinates of the recognition data. By analyzing the second theoretical coordinates and the visual recognition coordinates of the same recognition data, the recognition data that meets the screening conditions are marked, including: The n-2th visual recognition image p (n-2) The center coordinates of the identified data are determined as the second reference coordinates (x0', y0'); Determining the second theoretical coordinates (x3, y3) after the recognition data moves twice according to the second reference coordinates (x0', y0') and the preset movement distance; Calculating the second distance tolerance (Rx', Ry') of the recognition data according to the movement distance tolerance; Based on visual recognition image p (n) The visual recognition coordinates (x2, y2) of the recognition data are analyzed, and the second theoretical coordinates (x3, y3), the visual recognition coordinates (x2, y2) and the second distance tolerance (Rx', Ry') of the same recognition data are analyzed to determine whether the same recognition data meets the screening condition; When both |x3 - x2| < Rx and |y3 - y2| < Ry are satisfied, it is determined that the same recognition data meets the screening conditions.

8. The method for preventing weight loss in pipeline visual processing according to claim 7, characterized in that: Also including: The calculation method of the second distance tolerance (Rx', Ry') is expressed by the formula: ； 。

Images