Video Streaming Helmet Detection Algorithm Based On Feature Map Fusion And Faster RCNN

Safety helmet detection methods have many interference factors in the measurement site, and it is difficult to extract the characteristics of safety helmets. The accuracy of the existing convolutional neural network method for helmet detection needs to be improved. This paper proposes a video stream helmet detection algorithm based on feature map fusion and Faster RCNN. In the training of the detection model, the feature map fusion method is used to obtain a feature map with richer feature information, and then the feature map is used to train the detection model. In the detection process, the images are taken from the video streaming of the measurement site and the pre-trained model is used for detection. Experiments on the data set we created have verified the effectiveness of our proposed method. This method can effectively detect the safety helmets of electric workers in the measurement field. The detection accuracy of the method is 96%, which is higher than the existing detection method.