Design and implementation of intelligent construction site safety helmet real-time detection system based on Huawei Cloud and deep learning

Abstract

This project is a smart construction site safety helmet intelligent recognition system based on the Huawei Cloud platform, aiming to solve the problems of low efficiency, limited coverage, and delayed response of manual inspection in traditional construction site safety supervision. By receiving construction site videos uploaded by users, the system uses the target detection model trained on the Huawei Cloud Model Arts platform to realize real-time recognition of construction workers’ safety helmet wearing status (wearing/not wearing) and color attributes (red, yellow, blue, etc.) in video streams [1]. The core functions of the system include three modules: video analysis, target detection, and result visualization. First, frame extraction and preprocessing are performed on the uploaded video; then, the optimized detection model is called to accurately locate the personnel area and judge the safety helmet status and color; finally, the target is framed in real time through a bounding box, and the recognition result is labeled in box [2]. The platform deeply integrates artificial intelligence and cloud services, provides an efficient and low-cost solution for smart construction site safety management, effectively reduces the risk of safety accidents caused by non-standard safety helmet wearing, and has strong engineering application value [3].