An identification method and system for children's playground

Abstract

The invention provides an identification method and system for children's playgrounds. The clothing features are obtained by collecting and registering image data and performing feature extraction on the clothing data. The monitoring images and target coordinates are collected to form a monitoring data set, and the monitoring data set is registered according to The clothing features in the image data are matched, and the matching monitoring image is obtained as the target image, thereby obtaining the target coordinates of the camera that collects the target image, so as to realize the use of the clothing features in the monitoring image to monitor children in the playground. Long-distance efficient identification and precise positioning and tracking.

Description

technical field [0001] The invention belongs to the field of image processing technology and face recognition technology, and in particular relates to a recognition method and system for children's playgrounds. Background technique [0002] In large-scale public places, such as playgrounds, the identification and tracking of children is crucial to ensuring the safety of minors' lives and property. In modern children's playgrounds, cameras have been widely installed and applied. Multiple cameras in the amusement park are connected to the cloud server through a wireless network. Through the machine vision algorithm of artificial intelligence, it can identify and track the monitoring images in public places, so as to obtain the children's movement trajectory. Generally, target tracking algorithms based on deep learning require a large number of pre-training data sets, and the calculation time cost is relatively large, making it difficult to save resource costs. On the other h...

AI Technical Summary

Problems solved by technology

Generally, target tracking algorithms based on deep learning require a large amount of pre-training data sets, and the calculation time cost is relatively large, making it difficult to save resource costs.

On the other hand, due to the small size of the target child and the long distance from the surveillance camera equipment, the captured face images are small and blurred, making it difficult to identify the corresponding target

The patent document of Publication No. CN109145801B provides a face recognition method and a face recognition system to compare the extracted face features with a plurality of original pictures and on-site collected pictures respectively, and then to recognize the face again, and use the face picture as The new on-site collected pictures are stored in electronic devices, but it is still impossible to use the characteristics of large areas of clothing in the surveillance images, and it is impossible to achieve long-distance target recognition and tracking

Images