Machine-woven fishing net flaw real-time detection method based on artificial intelligence technology

Abstract

The invention discloses a machine-woven fishing net flaw real-time detection method based on an artificial intelligence technology. Mainly solves the problems of uneven effects and low efficiency when the flaws of the netting are checked by naked eyes in the conventional fishing net production. According to the machine-woven fishing net flaw real-time detection method based on the artificial intelligence technology, artificial intelligence and the deep learning technology are applied to traditional fishing net piece production, the combination of the new technology and the traditional industry is achieved, the competitiveness of the traditional industry is promoted to be improved, fusion innovation is achieved, and the domestic blank in the fishing net industry is filled up.

Description

Technical Field

[0001] The present invention relates to the field of fisheries, and in particular to a real-time detection method for defects in machine-woven fishing nets based on artificial intelligence technology. Background Art

[0002] In fishery production, the quality of fishing nets directly affects the operating efficiency of finished fishing nets. A major challenge in fishing net production is to be able to detect defects in fishing nets, such as broken mesh and missing mesh. In order to ensure the production of high-quality fishing nets, the industry has always relied on visual inspection of net defects, but the results are uneven and the efficiency is low. There is no automated detection system in the industry yet. With the rapid development of artificial intelligence technology and deep learning technology, they are rapidly penetrating and applied in various industries, providing new opportunities for improving the quality of traditional fishing net production. Summary of the invention

[0003] In order to overcome the shortcomings of the background technology, the present invention provides a real-time detection method for machine-woven fishing net defects based on artificial intelligence technology, which mainly solves the problem that the defects of the net are checked by naked eye in the current fishing net production, but the effect is uneven and the efficiency is low.

[0004] The technical solution of the present invention is: The real-time detection method of machine-woven fishing net defects based on artificial intelligence technology includes the following steps: Step 1: Standardized data set construction, collecting and annotating the image data of fishing net defects that meet and do not meet the standards, constructing a training data set and obtaining a deep learning model through training; Step 2: Establish a real-time detection system, collect real-time images through the image acquisition unit, analyze the collected fishing net images, identify all mesh areas of the fishing net fabric image, and analyze the different characteristics of normal meshes and defective meshes to determine defects; Step 3: Convert the image of the defective mesh obtained in step 2 into a grayscale image, continue to capture the image in the X-axis direction and / or Y-axis direction, and calculate the mesh density of the multiple meshes obtained and the total area of ​​the image, and compare it with the density of the standard mesh to determine whether it is at the edge of the fishing net. If not, mark it as an unqualified product image.

[0005] The image acquisition unit is a camera installed on the assembly line.

[0006] The deep learning model in step 1 is a model based on CNN convolutional neural network.

[0007] Step 2 also includes configuring a timestamp for each image.

[0008] Also included is a marking device for spraying a red dot on products determined to be unqualified.

[0009] The beneficial effects of the present invention are as follows: the present invention provides a real-time detection method for machine-woven fishing net defects based on artificial intelligence technology, applies artificial intelligence and deep learning technology to the production of traditional fishing nets, realizes the combination of new technologies and traditional industries, promotes the improvement of the competitiveness of traditional industries, realizes integrated innovation, and fills the domestic gap in the fishing net industry. DETAILED DESCRIPTION

[0010] The present invention is further described below: A real-time detection method for defects in machine-woven fishing nets based on artificial intelligence technology comprises the following steps: Step 1: Construct a standardized data set. Based on the standard "Woven Fishing Nets (GB_T 18673-2008)", collect and annotate fishing net defect image data that meets and does not meet the standard, build a training data set and obtain a deep learning model through training; collect fishing net images, classify and annotate them according to national standards, and design and train a deep learning model specifically for fishing net defect detection.

[0011] Step 2: Establish a real-time detection system. Collect real-time images through the image acquisition unit, analyze the collected fishing net images, identify all mesh areas of the fishing net fabric image, and analyze the different characteristics of normal meshes and defective meshes to determine defects; Embedded system adaptation, optimize the algorithm for specific embedded devices (such as Rockchip RV1126 and RK3588) to ensure real-time performance and efficiency; Step 3: Convert the image of the defective mesh obtained in step 2 into a grayscale image, continue to intercept the image in the X-axis direction and / or Y-axis direction, and calculate the mesh density of the multiple meshes obtained and the total area of ​​the image, and compare the density with the standard mesh to determine whether it is at the edge of the fishing net. If not, mark it as an unqualified product image. The mesh density of each fishing net is standardized and is within a certain threshold space. Continue to intercept the image of the suspected defective part in the X-axis and / or Y-axis direction, and calculate the mesh density of this image. If the mesh density is greater than 10% compared with the standard, it is determined to be an unqualified product.

[0012] The image acquisition unit is a camera installed on the production line, and can also be set at the scanning and packaging stage before storage.

[0013] The deep learning model in step 1 is a model based on CNN convolutional neural network.

[0014] Step 2 also includes assigning a timestamp to each image to facilitate searching.

[0015] The device also includes a marking device for spraying a red dot on the product determined to be unqualified. The marking device can be a spray gun, which can spray a certain amount of paint onto the unqualified product to facilitate searching.

[0016] In summary, the characteristics of this application are (1) technology application integration and innovation: applying artificial intelligence and deep learning technology to the production of traditional fishing nets, realizing the combination of new technologies and traditional industries, promoting the improvement of the competitiveness of traditional industries, and achieving integration and innovation.

[0017] (2) Applicable to system optimization of embedded devices, realizing highly efficient real-time detection technology, significantly reducing product defect rate, and reducing labor costs.

[0018] (3) This artificial intelligence technology fills a domestic gap in the fishing net industry.

[0019] In the description of the present invention, it is necessary to understand that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicating directions or positional relationships are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0020] The embodiments should not be regarded as limiting the present invention, but any improvements made based on the spirit of the present invention should be within the protection scope of the present invention.

Claims

1. A real-time detection method for defects in machine-woven fishing nets based on artificial intelligence technology, characterized in that: The following steps are included: Step 1: Standardized data set construction, collecting and annotating the image data of fishing net defects that meet and do not meet the standards, constructing a training data set and obtaining a deep learning model through training; Step 2: Establish a real-time detection system, collect real-time images through the image acquisition unit, analyze the collected fishing net images, identify all mesh areas of the fishing net fabric image, and analyze the different characteristics of normal meshes and defective meshes to determine defects; Step 3: Convert the image of the defective mesh obtained in step 2 into a grayscale image, continue to capture the image in the X-axis direction and / or Y-axis direction, and calculate the mesh density of the multiple meshes obtained and the total area of ​​the image, and compare it with the density of the standard mesh to determine whether it is at the edge of the fishing net. If not, mark it as an unqualified product image.

2. The method for real-time detection of defects in machine-woven fishing nets based on artificial intelligence technology according to claim 1, characterized in that: The image acquisition unit is a camera installed on the assembly line.

3. The method for real-time detection of defects in machine-woven fishing nets based on artificial intelligence technology according to claim 2, characterized in that: The deep learning model in step 1 is a model based on CNN convolutional neural network.

4. The method for real-time detection of defects in machine-woven fishing nets based on artificial intelligence technology according to claim 3 is characterized in that: Step 2 also includes configuring a timestamp for each image.

5. The method for real-time detection of defects in machine-woven fishing nets based on artificial intelligence technology according to claim 4 is characterized in that: Also included is a marking device for spraying a red dot on products determined to be unqualified.