Abstract
For video surveillance analysis, target detection algorithms are particularly important. Aiming at the problem of target detection in road traffic scenes, this paper proposes a non-motor vehicle target detection method that uses EdgeBoxes algorithm and deep learning fusion. It is inspired by deep learning target classification algorithm Fast R-CNN, and combines non-motor vehicle data samples in VOC format to integrate road traffic scenes. We use the EdgeBoxes algorithm to extract the target of the sample. It is recommended to construct an appropriate amount of region of interest, and inputs the network for iterative training with the sample. In the target detection of road traffic scenes, compared with the traditional method based on the fusion of the basic EdgeBoxes algorithm and the optimized Fast R-CNN method, the detection accuracy is slightly improved, the calculation is significantly reduced, and the algorithm has advantages in time complexity.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call