Forward vehicle detection based on deep convolution neural network

Aiming at vehicle target detection in real traffic scenarios, the representative Faster R-CNN framework of deep learning target classification algorithm is applied to transform the problem of target detection in scenarios into a two-class problem of target detection combined with the vehicle data set in ImageNet to detect and recognize vehicle targets. Compared with the other two target detection algorithms, R-CNN and Fast R-CNN, the front vehicle target detection algorithm based on Faster R-CNN has obvious advantages in detection accuracy and execution efficiency. The experimental results show that the recognition accuracy and speed of this method have been significantly improved.