Method and system for refrigerated shelf commodity statistics based on image processing technology

Abstract

The invention discloses an image processing technology-based method and system for statisticing products on shelves in refrigerators, and relates to the field of artificial intelligence, including: a terminal controller splicing a first image and a second image into a third image, and a cloud server dividing the third image It is a plurality of fourth images; then each fourth image is segmented to calculate the first total amount of goods placed in the first row in each storage compartment in the refrigerator; The total amount of goods is summed to obtain the second total amount of goods placed in the first row of the freezer and stored in the database. Stitching images according to the rotation invariance of the center of the circle can reduce the matching error of feature points and make the spliced ​​image better. Layering the spliced ​​image into the fourth image and finely segmenting the fourth image can accurately segment each Commodities placed in the first row in the storage compartment can improve the statistical efficiency, and can also improve the accuracy of commodity statistics, which is conducive to improving the management quality of refrigerator commodities.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence and digital image processing, and more specifically, to a method and system for statisticing commodities on refrigerator shelves based on image processing technology. Background technique [0002] Digital image processing technology is an interdisciplinary field. With the continuous development of computer science and technology, image processing and analysis has gradually formed its own scientific system, and new processing methods emerge in endlessly. widespread attention. [0003] When carrying out product statistics on refrigerator shelves, when the camera cannot capture the products on the entire shelf at close range, although a wide range of images can be obtained by extending the shooting distance, the products in the image may be relatively small, and mosaic phenomenon will appear after zooming in. The effect is poor. Although this problem can be solved by hardware...

AI Technical Summary

Problems solved by technology

[0003] When carrying out product statistics on refrigerator shelves, when the camera cannot capture the products on the entire shelf at close range, although a wide range of images can be obtained by extending the shooting distance, the products in the image may be relatively small, and mosaic phenomenon will appear after zooming in. The effect is poor. Although this problem can be solved by hardware equipment, hardware such as wide-angle lens is expensive, the operation is complicated, and it will cause image edge distortion. Therefore, image stitching and segmentation technology can be used for shelf commodity statistics.

[0004] In the prior art, there are many methods for image stitching and segmentation, but they all have certain defects. For example, the image stitching technology based on the Hough circle detection algorithm has a large error after stitching and registration; the rotation-translation invariance algorithm based on SURF , because there are problems such as reflection and high similarity in the freezer commodity image, the error is large when matching image feature points, resulting in a large error in splicing based on corresponding feature points; adaptive threshold segmentation algorithm, maximum inter-class threshold method, edge Neither the segmentation algorithm nor the region growing algorithm RGSA can accurately segment the commodities in the refrigerator shelves. Therefore, it is necessary to study a statistical method for refrigerator shelves based on image processing technology.

Images