1. Road recognition and motion target tracking based on infrared image
- Author
-
Qing Qing Wang, Li Guo Zheng, and Jia Nan Meng
- Subjects
Physics::Physics and Society ,Multi-mode optical fiber ,General Computer Science ,Line fitting ,business.industry ,Computer science ,ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION ,Boundary (topology) ,Kalman filter ,Tracking (particle physics) ,Noise ,Computer Science::Computer Vision and Pattern Recognition ,Trajectory ,Segmentation ,Computer vision ,Artificial intelligence ,Electrical and Electronic Engineering ,business - Abstract
Road recognition is a very important link in the safe driving of vehicles. In this paper, a region based road recognition method is proposed for infrared road images. Noise is eliminated by mean filtering and threshold segmentation. Then, the road boundary segments are obtained by the threshold exhaustive method, and the road boundary segments are obtained by the second-order differential operator. The road boundary segments obtained by the two methods are fused to obtain the candidate road boundary segments. Finally, the real road boundary is obtained by optimising the line fitting. For the identified road, the moving target tracking based on interacting multi-mode Kalman filter is adopted. The simulation results show that the interacting multimode Kalman filter can achieve better trajectory fitting, and the errors in X and Y directions before and after filtering are small. This shows the excellent performance of interacting multimode Kalman filter in moving target detection.
- Published
- 2021