CN-122016585-A - Particle size detection method and device, electronic equipment and medium

CN122016585ACN 122016585 ACN122016585 ACN 122016585ACN-122016585-A

Abstract

The invention relates to the technical field of particle size detection, in particular to a particle size detection method, a particle size detection device, electronic equipment and a medium. The method can automatically determine the proper clustering threshold range by adopting a mode of screening and then counting, does not need to carry out a large number of parameter tuning aiming at a specific scene, and has strong generalization capability. And each cluster generates a representative detection point of a particle area, and the representative detection point is used as a prompt point to be input into a segmentation model for image segmentation to obtain the particle contour of the corresponding particle area, so that the segmentation accuracy is high. And finally, accurately acquiring the particle size of the particles according to the particle profile, and ensuring the accuracy and reliability of the measurement result. The invention has the capability of processing a plurality of particles in the material image at one time, shortens the detection time and can meet the requirement of quick detection of a production line.

Inventors

KONG YIFAN
ZHANG YANCHEN
XU JIAWEN

Assignees

合肥国轩高科动力能源有限公司

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (10)

1. A particle size detection method, comprising: Acquiring a material image to be detected; Preprocessing the material image, and screening particle pixels from the preprocessed material image; Determining a clustering threshold range according to the statistical distribution condition of the pixel values of the particle pixels; Performing cluster analysis on pixel points with pixel values within the cluster threshold range in the particle pixels to obtain a plurality of cluster clusters, wherein each cluster corresponds to one particle region; and for each cluster, generating a representative detection point of the particle area, inputting the representative detection point as a prompt point into a segmentation model for image segmentation to obtain a particle contour in the particle area, and determining the particle size according to the particle contour.
2. The method of claim 1, wherein screening out particle pixels from the pre-processed material image comprises: and filtering background pixels in the preprocessed material image based on a preset pixel value threshold value to retain particle pixels.
3. The method for detecting particle size according to claim 1, wherein the statistical distribution comprises a normal distribution; determining a cluster threshold range from the statistical distribution of pixel values of the particle pixels comprises: And determining a clustering threshold range according to the mean value and the standard deviation of the normal distribution condition of the pixel values of the particle pixels.
4. The method according to claim 1, wherein performing cluster analysis on pixel points having pixel values within the cluster threshold range among the particle pixels to obtain a plurality of clusters comprises: For each pixel point of which the pixel value is in the clustering threshold range in the particle pixels, determining the density value of the pixel point according to the number of the pixel points belonging to the clustering threshold range in the neighborhood of the pixel point; and gathering the pixel points which are adjacent in position and have the density value higher than a preset density threshold value to obtain a cluster.
5. The method according to claim 4, wherein the representative detection points of the generated particle region include: For each pixel point of the cluster, weighting the space coordinates of the pixel point by using the density value to obtain density weighted coordinates; and calculating the density centroid of the particle area according to the density weighted coordinates of all the pixel points to serve as the representative detection point.
6. The method according to claim 1, wherein image-dividing the representative detection point input division model to obtain a particle contour comprises: inputting the representative detection points into a SAM model to divide the representative detection points into preliminary contours of particles; and carrying out post-processing algorithm optimization on the preliminary profile to finally obtain a particle profile.
7. The method of claim 1, wherein determining the particle size based on the particle profile comprises calculating a diameter of a smallest circumscribed circle of the particle profile as the particle size of the particle.
8. A particle size detection device, comprising: The acquisition module is used for acquiring the material image to be detected; The preprocessing and screening module is used for preprocessing the material image and screening particle pixels from the preprocessed material image; the range determining module is used for determining a clustering threshold range according to the statistical distribution condition of the pixel values of the particle pixels; the clustering module is used for carrying out cluster analysis on pixel points with pixel values within the cluster threshold range in the particle pixels to obtain a plurality of cluster clusters, and each cluster corresponds to one particle region; The particle size determining module is used for generating a representative detection point of the particle area for each cluster, inputting the representative detection point as a prompt point into the segmentation model for image segmentation to obtain a particle contour in the particle area, and determining the particle size according to the particle contour.
9. An electronic device comprising a processor and a storage medium for storing instructions, the processor being operative in accordance with the instructions to perform the steps of the method as claimed in any one of claims 1-7.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the method according to any one of claims 1-7.

Description

Particle size detection method and device, electronic equipment and medium Technical Field The invention relates to the technical field of particle size detection, in particular to a particle size detection method, a particle size detection device, electronic equipment and a medium. Background Particle size detection based on material pictures plays an extremely important role in the field of battery cell production, and is directly related to quality control and product performance of battery materials. In the manufacturing process of lithium ion batteries, the particle size distribution of the anode and cathode materials is one of the key factors affecting the battery capacity, cycle life and safety performance. The traditional method mainly relies on manual measurement on a microscope image by using a staff gauge, and the method is time-consuming and labor-consuming, has high measurement accuracy greatly influenced by experience of operators, has poor repeatability and is difficult to meet the requirements of modern production lines on efficient and accurate detection. Traditional image processing algorithms (such as methods of edge detection, threshold segmentation, region growing and the like) are poor in performance when processing noise interference, illumination change, adhesion particles and other imaging scenes with interference, and often cannot accurately segment single particles. The method generally needs to perform a large amount of parameter tuning aiming at a specific scene, has poor generalization capability, is difficult to process a large amount of particles in an image at one time, has low detection efficiency, and cannot meet the requirement of real-time online detection. Disclosure of Invention The invention aims to provide a particle size detection method, a device, electronic equipment and a medium, which realize high-efficiency and accurate particle size detection by combining a clustering algorithm and an image segmentation algorithm, can accurately segment single particles, do not need to carry out a large amount of parameter tuning for a specific scene, can treat a plurality of particles in a material image at one time, and remarkably improve the quality detection efficiency and accuracy in the production process of an electric core. In a first aspect, the present invention provides a particle size detection method comprising: Acquiring a material image to be detected; Preprocessing the material image, and screening particle pixels from the preprocessed material image; Determining a clustering threshold range according to the statistical distribution condition of the pixel values of the particle pixels; Performing cluster analysis on pixel points with pixel values within the cluster threshold range in the particle pixels to obtain a plurality of cluster clusters, wherein each cluster corresponds to one particle region; and for each cluster, generating a representative detection point of the particle area, inputting the representative detection point as a prompt point into a segmentation model for image segmentation to obtain a particle contour in the particle area, and determining the particle size according to the particle contour. The method comprises the steps of preprocessing a material image to be detected (comprising noise removal, contrast enhancement and image standardization), screening particle pixels from the preprocessed material image, and determining a clustering threshold range according to the statistical distribution condition of the pixel values of the particle pixels. And then carrying out cluster analysis on pixel points with pixel values within the cluster threshold range in the particle pixels to obtain a plurality of cluster clusters, wherein each cluster corresponds to a particle area, so that adhered particles can be effectively distinguished, the accurate positioning of single particles is realized, each cluster generates a representative detection point of a particle area, the representative detection point is used as a prompt point to be input into a segmentation model for image segmentation to obtain the particle contour of the corresponding particle area, and the segmentation accuracy is high. And finally, accurately acquiring the particle size according to the particle profile, and ensuring the accuracy and reliability of the measurement result. The invention has the capability of processing a plurality of particles in the material image at one time, improves the detection efficiency by a plurality of times compared with the traditional method, greatly shortens the detection time and meets the requirement of quick detection of a production line. Optionally, screening the particle pixels from the preprocessed material image comprises: and filtering background pixels in the preprocessed material image based on a preset pixel value threshold value to retain particle pixels. The invention filters background pixels based on the preset pixel value thresh