Automatic Identification of Traffic Accidents based on Intelligent Identification

The traffic accident process system is complex, by extracting the traffic accident information, the real-time state monitoring of the traffic accident image can be realized. A method of traffic accident information extraction and automatic recognition is proposed based on depth neural network and intelligent image analysis. The remote sensing method is used to collect the traffic accident image, the traffic accident remote sensing image is adaptively de-noised, and the block feature matching method is used to deal with the traffic accident image block fusion. The feature extraction of traffic accident state is carried out in the subspace component of traffic accident image, and the image feature classification of traffic accident is processed by using depth neural network classifier to realize image feedback recognition of traffic accident state feature quantity. A large number of traffic video images are selected for experiments. The simulation results show that the proposed method is more intelligent and accurate.