Moving target detection and track prediction method based on image processing

Abstract

The invention discloses a moving target detection and track prediction method based on image processing, and the method comprises the steps: carrying out the image processing, target detection and positioning, and target feature extraction of a target when carrying out the position prediction of a moving target in a video image at the next moment, wherein the part is composed of a main body detection part and a detail detection part, carrying out target area marking on a final foreground image after target detection is carried out, and then extracting and quantifying feature; improving the noise of the Kalman filtering algorithm by using a particle swarm optimization algorithm; and taking the quantized values extracted from the Kalman predicted position coordinates and the target featuresas input parameters of the Elman neural network, taking the corresponding real position coordinates as output values, improving the parameters of the Elman neural network by using a particle swarm optimization algorithm, and finally obtaining the predicted position of the moving target through network training. Detection and prediction errors of the method are small, and the precision of the method is greatly improved compared with other methods.

Description

technical field [0001] The invention belongs to the field of image processing in computer technology, and in particular relates to a method for detecting and predicting a moving target based on image processing. Background technique [0002] In recent years, with the continuous development of the field of computer vision, the research on video and image processing is being further developed and applied. As far as the video image is concerned, by obtaining the moving data of the target in a certain time series, and fully mining its internal characteristics after image processing, the relevant methods can be used to track and locate the moving target and predict the possible future trajectory. Compared with a single image Static targets in , can discover more meaningful and valuable information, and have a wider range of applications. [0003] When predicting the position of a moving target in a video image at the next moment, it is first necessary to perform image processing...

AI Technical Summary

Problems solved by technology

At present, the recognized target detection methods include background difference method, frame difference method, optical flow method, and Gaussian mixture model, but the above methods are often polluted by noise when detecting targets. With some difficulty, it affects the foreground detection rate

Then, when predicting the trajectory of the target, although the existing Kalman filter algorithm and Bayesian filter algorithm can predict well, they are often affected by improper random noise points on the system, thus affecting the prediction efficiency and accuracy. degree, resulting in the predicted position of the next moment is too far from the real position

Images