A Pixel Coordinate Trajectory Prediction Method for Moving Objects in Video Images

Abstract

The invention discloses a method for predicting the pixel coordinate trajectory of a moving target in a video image. Based on the historical pixel trajectory information of the moving target obtained from the video image, the method proposes a fractional model fitting trajectory formula based on the camera imaging principle. , the fractional model coefficients can be obtained by establishing a system of overdetermined equations based on historical trajectories, thereby determining the relationship between the pixel coordinates and time of the uniform linear motion target in the actual space in the video image, and finally accurately predicting the pixel coordinates of the target in the future. . This invention does not need to provide camera parameters or any information on the actual road surface in advance. It only needs to rely on the historical pixel movement trajectory and corresponding time interval information of the uniformly moving target in the actual space in the video image to predict its image position in the future. , easy to operate, practical and highly accurate.

Description

Technical field: [0001] The invention belongs to the technical field of video image signal processing, and specifically relates to a prediction method for determining the corresponding relationship between pixel coordinates and time of a moving object by using historical pixel track information of the moving object. Background technique: [0002] With the development of the economy and the improvement of people's living standards, the level of motorization in China has increased rapidly, and there have been more and more traffic jams and accidents. Therefore, the supervision of vehicles on the road has brought enormous pressure to the traffic police. As a technology that can effectively solve the appeal problem, intelligent transportation has been paid more and more attention by people. [0003] As the resolution of current traffic monitoring cameras is getting higher and higher, the monitoring range is also getting bigger and bigger. In the real-time video images on the roa...

AI Technical Summary

Problems solved by technology

[0004] At present, the more accurate prediction method is to establish the relationship between the pixel distance and the actual distance based on data such as camera parameters, erection height, attitude, and road distance, and then deduce the relationship between the pixel movement distance and time of the target through the law of uniform motion in the real world. relationship, but for traffic cameras that need to be widely used, in actual operation, some data have large errors or even cannot be obtained at all.

Images