Computer vision-based express parcel violent sorting identification method

Abstract

The present invention discloses a computer vision-based express parcel violent sorting identification method. The method comprises the following steps of: a depth camera-based pose estimation: a human pose estimation problem is converted into a problem of classifying depth image pixels captured by a depth camera, and human body pose estimation is realized by using a random forest method; human body three-dimensional pose relative spatial-temporal feature extraction: relative spatial-temporal positions of geometric elements of points, lines and surfaces formed by joints in three-dimensional poses and the measures of the change of the relative spatial-temporal positions are extracted and are adopted as the feature representations of the poses; and recurrent neural network-based violent sorting identification: modeling training is performed on the pose spatial-temporal relative features which are continuous in time and are extracted from the human body three-dimensional poses through a long and short memory model (LSTM), so that the identification of express parcel violent sorting behaviors can be realized.

Description

technical field [0001] The invention belongs to the technical field of computer vision and pattern recognition, and relates to a computer vision-based express violent sorting and recognition method. Background technique [0002] With the increasing demand for express delivery services in new service industries such as e-commerce and online shopping, my country's express delivery industry has shown a trend of rapid development. Express delivery has become a livelihood service that connects hundreds of millions of merchants and the general public. It plays an increasingly important role in promoting consumption upgrading. Compared with previous years, the service quality and sorting efficiency of my country's express delivery industry are constantly improving, but industry problems such as "violent sorting" are still emerging one after another. Open Bing or Baidu search, and there are hundreds of thousands of web pages related to violent sorting. The chaos of "violent sorting"...

AI Technical Summary

Problems solved by technology

[0007] Although human behavior recognition based on deep learning has made some progress, however, human behavior recognition based on 2D video still faces many challenges, including the difference of inter-class and intra-class data and scene changes. The person itself is a complex non-rigid object. Human action recognition is affected by individual differences in human appearance, posture, movement, clothing, perspective changes and camera movements, lighting changes, occlusions and complex backgrounds, while 2D videos only get 3D A projection of space, which causes great difficulties for the underlying action recognition, and it is difficult to meet the needs of courier violence sorting and recognition in actual complex scenes

[0008] In summary, the existing human behavior recognition methods based on computer vision are difficult to meet the needs of express violent sorting and recognition

Images