Research on Mask Wearing Detection Algorithm in Complex Environment
Under the requirements of epidemic prevention and control,mask wearing detection is affected by interference fac-tors,complex environment and other aspects,so efficient wearing detection is of great significance.To solve the above problems,an improved algorithm based on YOLOv5 is proposed.Firstly,mask data set is preprocessed in complex scenes,including frequency domain image enhancement algorithm,Mosaic and Mix up image fusion algorithm.Secondly,change the network model CSP mod-ule of YOLOv5 into CSP1-1.Some CBL modules are replaced with CBS modules.In addition,three CBAM attention mechanism modules were added to improve its detection accuracy by 1.9%on the basis of maintaining the original model,enabling real-time and accurate mask wearing detection for pedestrians.
complex environmentmask wearing testYOLOv5data to enhanceattentional mechanism
国家科技期刊平台
NSTL
万方数据
