Title: Study on abnormal behaviour recognition of substation construction site based on dynamic feature fusion

Authors: Yun Zhao; Ziwen Cai; Yuxin Lu; Wei Cao

Addresses: China Southern Power Grid Research Institute Co., Ltd., Guangzhou, 51066, Guangdong, China ' China Southern Power Grid Research Institute Co., Ltd., Guangzhou, 51066, Guangdong, China ' China Southern Power Grid Research Institute Co., Ltd., Guangzhou, 51066, Guangdong, China ' China Southern Power Grid Research Institute Co., Ltd., Guangzhou, 51066, Guangdong, China

Abstract: This paper proposes an abnormal behaviour recognition method of substation construction site based on dynamic feature fusion. Kinect camera is used to collect the video of substation construction site, and the key frames of the video are extracted by frame difference method. The obtained image is greyscaled and Gaussian filtered, and the image target is detected by convolutional neural network. The dynamic features of the target, such as texture features, motion features and shape features, are fused, and the abnormal behaviour is recognised and abnormal behaviour is recognised by combining the dynamic features with a multi-layer perceptron. Experimental results indicate that the feature extraction accuracy of the proposed method ranges from 95.27% to 97.48%, with a maximum accuracy rate of 98.63%, the recognition time varies from 0.16s to 0.86s.

Keywords: dynamic feature fusion; substation; construction site; abnormal behaviour recognition; texture features; texture features; shape features.

DOI: 10.1504/IJCAT.2026.153109

International Journal of Computer Applications in Technology, 2026 Vol.78 No.3, pp.194 - 203

Received: 04 Dec 2024
Accepted: 22 Apr 2025
Published online: 22 Apr 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article