Method for judging target identity of articles in refrigerator, refrigerator and computer storage medium

Abstract

The invention provides a method for judging the target identity of articles in a refrigerator, the refrigerator and a computer storage medium, and the method comprises the steps: obtaining a plurality of storage space pictures in the refrigerator, namely a first image to an nth image; article information in the first image is detected, target detection frames are generated at the positions of all the articles, the target detection frames B1 to Bn are the target detection frames B1 to Bn, and ID information is assigned to the articles in the target detection frames and corresponds to targets 1 to n. Article information in the second image is detected. And comparing the overlap ratio between the target detection frame of each article in the second image and the target detection frame B1 to obtain a target detection frame Bmax with the highest overlap ratio, detecting the similarity between the target detection frame Bmax and the target detection frame Bmax, if the similarity meets the lowest requirement, assigning the article in the target detection frame Bmax as a target 1, otherwise, re-checking and deleting the information of the target 1. And repeating the steps to update the image information. According to the invention, the information and position change of each article can be detected and tracked with high accuracy in a section of continuous image.

Description

technical field [0001] The invention relates to the field of refrigeration devices, in particular to a method for judging object identity of items in a refrigerator, a refrigerator and a computer storage medium. Background technique [0002] With the development of intelligent household appliances, the identification of internal items has become a necessary function of only refrigerators. Usually, in order to realize the identification and tracking of the items inside the refrigerator, to detect the storage date and storage time of the ingredients, etc., one or more cameras are installed in the internal storage space of the refrigerator for shooting items, and the photos are processed and generated. Identify the results. [0003] However, in the process of image detection and recognition, the existing Sort and deepsort algorithms are not suitable for the target tracking of items stored on the refrigerator bottle holder, because after the refrigerator door is opened, various...

AI Technical Summary

Problems solved by technology

[0003] However, in the process of image detection and recognition, the existing Sort and deepsort algorithms are not suitable for object tracking of items stored on refrigerator bottle holders, because various light changes and the occlusion of objects by the human body will occur after the refrigerator door is opened. During the process of opening the door, objects may shake in a wide range, resulting in the target not being detected for a long time, and even the target moves when the target is covered by the human hand for a long time, resulting in the loss of the target, or the rapid switching of the target ID, which seriously affects the user experience.

Images