Feature extraction method for identifying and tracking moving target

Abstract

The invention provides a feature extraction method for identifying and tracking a moving target. The method comprises the following steps: inputting an image and preprocessing the input image; using aGaussian kernel convolution function to generate a scale space pyramid; detecting feature points by adopting a Hessian matrix, roughly positioning the feature points, accurately positioning the feature points, extracting the main direction and the main curvature of the geometric surface of the image, calculating the gradient and the direction of the feature points according to the main curvature,and counting and labeling the obtained gradient amplitude and direction information by utilizing a method similar to HOG. According to the method, through using the Gaussian convolution function to construct the scale space, the scale invariant characteristic of the extracted feature points can be improved. The feature points of accurate positioning can be accurately acquired, gradient amplitudeand direction information of the acquired feature points can be guaranteed, good robustness on illumination can be realized, and good robustness on illumination, rotation, scale, visual angle, non-rigid deformation and other changes is realized.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a feature extraction method for moving target recognition and tracking. Background technique [0002] Artificial intelligence (AI) is the hottest research field in the world today, and computer vision, as an important research hotspot in artificial intelligence technology, plays a vital role in promoting the development and progress of artificial intelligence. At present, computer vision has Widely used in target recognition and tracking, scene understanding, medical image analysis and product quality inspection and other fields; [0003] Before identifying and tracking the moving target, it is first necessary to extract the features of the target and use the extracted features as a matching template to complete the identification and tracking of the target. In a real environment, the moving target will enter from a bright place under natural light In the shadow, the im...

AI Technical Summary

Problems solved by technology

[0003] Before identifying and tracking the moving target, it is first necessary to extract the features of the target and use the extracted features as a matching template to complete the identification and tracking of the target. In a real environment, the moving target will enter from a bright place under natural light In the shadow, the imaging results of the image are very different from those in the bright area, which seriously affects the image acquisition quality and finally leads to the wrong identification and tracking loss of the moving target. When multiple imaging conditions occur at the same time, the traditional moving target The feature extraction method for recognition and tracking is only robust to one or two of illumination, rotation, scale, and non-rigid deformation, but when these four external conditions change at the same time, it will seriously affect the moving target. image quality and recognition accuracy, especially in the case of non-rigid deformation, such as when the moving target is a person, people often rise, squat, run, jump or bend over when they are in motion, which will lead to During recognition, there is a big difference between the feature matching template and false matching occurs. In addition, during the movement process, due to the influence of the shooting angle of the video acquisition device, the target that may be collected is the result of imaging under different viewing angles, which provides automatic recognition of the target. and tracking have brought a great impact, which seriously restricts the effect of moving target recognition and tracking in practical applications. Therefore, the present invention proposes a feature extraction method for moving target recognition and tracking to solve the problems of existing technologies. inadequacies in

Images