CN-122023869-A - Sea sedge image capturing method and system based on multichannel conveyor belt

CN122023869ACN 122023869 ACN122023869 ACN 122023869ACN-122023869-A

Abstract

The invention provides a sea sedge image capturing method and system based on a multichannel conveyor belt, and belongs to the technical field of computer vision and food industry automatic detection. The system realizes high-quality capturing and processing of sea weed images by constructing a closed imaging environment, optimizing acquisition parameters and combining a lightweight contour recognition algorithm (B-CR-AF), a state transition batch tracking mechanism and a fusion denoising method (G-SFFT). The method solves the technical problems of illumination interference, contour redundancy, batch blurring, complex noise and the like in the sea weed image acquisition, has high precision, high instantaneity and strong robustness, and is suitable for industrial continuous production environments.

Inventors

YANG XU
Wang Hanran
SU QIANG
Ji Langpo
CHEN YONG
LIU DAXIN
XING HONGYAN
ZHU FA
WANG HUI

Assignees

南通大学

Dates

Publication Date: 20260512
Application Date: 20251223

Claims (9)

1. A sea sedge image capturing method based on a multichannel conveyor belt is characterized by comprising the following steps: s1, image acquisition and parameter optimization; s2, preprocessing and contour recognition; s3, batch tracking and image optimization; s4, denoising and optimizing output of the image.
2. The multi-channel conveyor-based sea weed image capturing method according to claim 1, wherein in S1, the image acquisition and parameter optimization comprises the steps of: The method comprises the steps of S11, optimizing camera parameters based on a secondary fuzzy comprehensive evaluation model, wherein the secondary fuzzy comprehensive evaluation model is used for determining index weights of all levels according to expert experience by constructing a defect characteristic characterization index system and subjective scoring grade standard; s12, establishing a mathematical model containing frame rate constraints and captured picture constraints so as to determine an optimal image capturing time interval; the frame rate constraint is: ; Wherein the method comprises the steps of For the frame rate of the camera, Is the minimum capture time interval; the captured picture constraint is: ; Wherein the method comprises the steps of For the length of the sea sedge, Is the interval between the front sea sedge and the back sea sedge, Capturing a range length for the camera; S13, combining channel rate And safety factor of Calculating a capture time interval: (3)。
3. A multi-channel conveyor belt-based sea weed image capturing method according to claim 1, wherein in S2, the preprocessing and contour recognition adopts a B-CR-AF algorithm, comprising the steps of: S21, image preprocessing, namely sequentially performing gray level conversion, gaussian filtering, global binarization based on a gray level histogram common valley point and morphological opening operation; s22, contour extraction and screening, namely, by means of binary images Extracting all closed contours by function, calculating area of each contour By performing statistical analysis of area pixel values of sea weed image, the sea weed image is based on normal distribution Criterion set area threshold range And screening the contours, and reserving the contours with the area within the range as effective sea weed contours.
4. A multi-channel conveyor belt-based sea weed image capturing method according to claim 1, wherein in S3, batch tracking and image optimization comprises: Set frame index Defining t frame complete sea sedge contour identification state ; (4); Batch signal triggering rules based on Respectively define batch start signals And end of batch signal ; (5); If present And (3) with So that And is also provided with The set of frames contained in the batch is (6); A complete batch is made up of a set of consecutive frames contained by a one-time start signal and an immediately following next end signal; Cutting the images to obtain the same sea weed image set ; For candidate images The variance of the Laplace feature graph is defined as the sharpness score : (7); Selecting the image with the highest definition score for the candidate image set ) As the sole representative output of this seaweed: (8)。
5. the multi-channel conveyor belt-based sea weed image capturing method according to claim 1, wherein in S4, the image denoising and optimizing output adopts a G-SFFT algorithm, comprising the steps of: S41, SVD deblurring treatment, namely, matrix of gray image Singular value decomposition and reconstruction are carried out through singular value duty ratio parameters Controlling the deblurring intensity to obtain a deblurred image: (9); For deblurred images Performing a fast fourier transform using a cut-off frequency Low pass filter of (2) Performing frequency domain filtering, and performing inverse transformation to obtain a denoising image: (10); s42, adopting improved genetic algorithm to combine parameters Performing global optimization; the fitness function is four indexes of a structural similarity index, a peak signal-to-noise ratio, an image information entropy and a mean square error: (11) Wherein, the 、、 And Respectively normalizing the corresponding indexes; , , , the weight coefficient corresponding to each index is determined by a hierarchical analysis method.
6. The system is characterized by comprising an image acquisition module for constructing a closed imaging environment and performing parameter optimization, a contour recognition module for performing a B-CR-AF algorithm to complete preprocessing and contour recognition, a batch tracking module for performing batch definition and image optimization based on a state transition mechanism, and an image denoising module for performing a G-SFFT algorithm to achieve image denoising and parameter self-adaptive optimization, wherein the system finally outputs high-quality sea weed images for subsequent defect detection through the cooperative work of the modules.
7. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the computer program is executed to implement the steps of the method according to any of claims 1 to 6.
8. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program is configured to implement the steps of the method of any one of claims 1 to 6 when called by a processor.
9. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of any of claims 1 to 6.

Description

Sea sedge image capturing method and system based on multichannel conveyor belt Technical Field The invention relates to the technical field of computer vision and food industry automatic detection, in particular to a high-performance image capturing method and system for sea sedge continuous production quality control based on software and hardware collaborative design. Background Sea sedge is a popular food, and the automatic requirement of production quality control is urgent. At present, the traditional manual detection method has the problems of strong subjectivity, low efficiency, easy environmental interference and the like, and is difficult to meet the requirement of high-speed continuous production. Although the detection technology based on computer vision improves the detection efficiency to a certain extent, the following problems still face in the sea weed image acquisition process. First, imaging environmental interference problems. The seaweed has semi-transparent characteristic, imaging is easily influenced by ambient light, and image contrast is unstable; second, the contours identify redundancy problems. The traditional edge detection algorithm is sensitive to sea weed surface textures, a large number of false contours are easy to generate, and recognition accuracy and instantaneity are affected; third, batch definition is a difficult problem. Under a multi-channel high-speed transmission scene, the sea weed units are similar in appearance, have fuzzy boundaries, and are difficult to accurately define batch starting and stopping; fourth, image noise is a complex problem. Light transmission artifacts, motion blur and high-frequency noise exist in the acquisition process, and noise suppression and detail preservation are difficult to be achieved by a single denoising method. How to solve the above problems is a subject of the present invention. Disclosure of Invention The invention aims to provide a multi-channel conveyor belt-based sea weed image capturing method and system based on fusion contour recognition, batch tracking and denoising optimization, and aims to solve the problem of image capturing in multi-channel sea weed continuous production. In order to achieve the aim of the invention, the invention adopts the technical scheme that the sea sedge image capturing method based on the multichannel conveyor belt comprises four steps of image acquisition and parameter optimizing, preprocessing and contour identifying, batch tracking and image optimizing, and image denoising and optimizing output: Wherein, the image acquisition and parameter optimizing step further comprises: A closed static acquisition device is deployed to stabilize the imaging environment. And carrying out definition parameter optimization based on a secondary fuzzy comprehensive evaluation model, and determining the optimal definition parameter configuration by constructing a defect characteristic characterization index system and subjective scoring grade standard, determining weights according to expert experience, and carrying out fuzzy comprehensive evaluation on each group of camera parameter configuration. And establishing a mathematical model containing frame rate constraints and captured picture constraints to perform integrity parameter optimization. The global shutter camera frame rate employed isCalculating a minimum capture time interval according to equation (1): Assume that the sea sedge is of a size of lengthMultiplying by widthGlobal shutter capture range length. In order to avoid the simultaneous appearance of adjacent sea sedge before and after a single frame image and reduce the complexity of image processing, the interval parameters of the sea sedge before and after are introducedThe requirements are as follows: According to equation (2), the minimum interval is calculated 。 Assuming that the rate of the sea weed processing channel is. To cope with channel rate fluctuation and sea weed position deviation, a safety factor is introducedThe method comprises the following steps of: obtaining a capturing time interval by using (3) To further select the capture time interval. Preprocessing the image by gray level conversion, gaussian filtering, global binarization and morphological open operation; further, a lightweight contour recognition algorithm (B-CR-AF) is used to recognize contours. By means ofThe function directly searches and extracts closed connected boundaries on the binarized image, and a complete contour set is preliminarily obtained. For a set of by pointsFormed as a closed contour with pixel areaDefined as the total number of pixels enclosed by the contour, as obtained in (4): statistical analysis of area pixel values of binarized sea weed image based on normal distribution Criterion setting a reasonable area threshold rangeContour screening is performed. The state transition mechanism performs batch definition. And then filtering the sea sedge image through image cutting and definition scor