Multi-size shelf commodity detection method

Abstract

The invention discloses a multi-size shelf commodity detection method. The method comprises the following steps: (1) obtaining a to-be-detected shelf image; (2) preprocessing the shelf images in the step (1), detecting shelf edge positions in the images through edge detection, Hough transform and linear screening, segmenting the images with excessive shelf layers, and correcting distortion of theimages with inclined shelves; (3) inputting the image obtained in the step (2) into a feature extraction network to obtain five layers of feature maps with different depths; (4) performing feature fusion on the obtained feature maps of each layer; (5) performing regional nomination on the feature map through a regional nomination network to obtain candidate boxes; and (6) further correcting the candidate box by using the target detection network and reasoning the type and accurate coordinate position of the to-be-detected commodity. The multi-size shelf commodity detection method based on image segmentation and the deep neural network is high in detection precision.

Description

technical field [0001] The invention relates to the field of target detection in computer vision, in particular to a commodity detection method. Background technique [0002] In order to better understand the market and make decisions about market launch and management, companies that produce consumer goods need to frequently check offline channel stores and investigate the distribution rate, number of rows and share of their products on store shelves, etc. . [0003] Usually, when an enterprise conducts a survey on store sales, it needs to arrange special staff to go to a designated store to make statistics on the display of goods on the store shelves, and to feed back the statistical results to the company. This manual verification method has many disadvantages. It not only consumes a lot of manpower, but also has low efficiency and untimely information update. Enterprises are often limited to cost and can only do sample surveys, unable to obtain accurate and comprehensi...

AI Technical Summary

Problems solved by technology

However, in the actual application scenario of product detection, there are a series of problems

The size of most commodities is relatively small. At the same time, because the shelves of each store are placed differently, the distance and angle between the camera equipment and the shelves cannot be kept consistent when shooting images. is inconsistent

However, the current popular target detection method based on deep neural network has relatively low detection accuracy when encountering the above situations, especially when the target object size is too small, it is often impossible to detect the target.

Images