Research On Vehicle Color Recognition Based On Deep Learning

In the intelligent traffic management system, in order to make up for the deficiency of the license plate recognition in the complex real scene,the color as an important expression in the visual information makes the vehicle color recognition become an important research topic. However,in complex scenes, vehicle color recognition faces many challenges. For example, it is difficult to distinguish between multiple color categories,While the inherent color of the body is susceptible to weather and lighting.To solve the above problems, this thesis studied color feature extraction and vehicle color recognition based on deep learning. In this thesis, a multiscale comprehensive feature fusion convolutional neural network based on residual learning (MCFF-CNN) is proposed to extract the deep color features. Firstly, MCFF-CNN reconstructs the learning function of the network layer through residual mapping. Secondly, the output features of the network layer with the same depth and different sizes are merged to realize the multi-scale fusion of image features.Finally, the output features of different depth network layers are merged to realize the fusion of local features and global features. The MCFF-CNN network design not only considers the efficiency and practicability of the computation,but also lowers the memory utilization rate and improves the learning performance of the network. The MCFF-CNN network uses the MCFF-CNN network to extract the color characteristics of the vehicle to show good robustness.In order to evaluate the effectiveness of the proposed vehicle color recognition system, we carried out a large number of experimental verification. The experimental results show that our method has better performance than the current method of vehicle color recognition.