| In computer vision,feature description and image matching is a important branch,which is widely applied in many fields such as object recognition,object tracking,satellite remote sensing,image interpretation etc.In this paper,we study the existing traditional image matching algorithm at home and aboard further and apply in vehicle tracking in the field of the intelligent transportation system(ITS).Taking the complex background factors into account,such as shadow,light,occlusion,we give consideration to both real-time and accuracy and put forward a adaptive system of vehicle detection and tracking in the real road environment.The main work is:(1)In the vehicle detection stage,vehicles occlusion will cause the loss of vehicle visual information,seriously affecting the accuracy of vehicle detection.In this paper,we proposed a vehicle detection method based on part-based model for complex scene.The image is divided into two regions according to common occlusion during model training.The two parts images are modeled and trained.During the testing process,the training model is used to detect the two regions of the vehicle through iteration,and the position and scale of learning are combined to make the training.(2)In the vehicle tracking stage,the accuracy of tracking and the real-time algorithm are the main concerns.In this paper,we proposed a vehicle tracking method based on SURF feature hash matching.After the SURF feature descriptors of the detection region image and the vehicle image to be tracked are extracted,the local sensitive hash algorithm is used to match.The matching set of feature points is matched and refined by using the global information that the included angles of the neighboring feature points are consistent,and the tracking accuracy and the real-time of the algorithm are improved by the regional tone value constraint.(3)The modular design is used in this paper.Based on the proposed vehicle detection and vehicle tracking algorithms,complex vehicle detection and tracking system in complex scenes is implemented on the PC. |
