Compressed sensing-based multi-view target tracking and 3D target reconstruction system and method

Abstract

The invention belongs to the field of image processing. In order to realize the purposes of not only reducing complexity at an encoding end and reducing transmission data bandwidth, but also improving adaptability and signal noise immunity of the algorithm and having robustness to background changes, the invention discloses a compressed sensing-based multi-view target tracking and 3D target reconstruction system and method, the technical scheme comprises the following steps: enabling cameras to be distributed at positions capable of observing the majority of observation areas, wherein the observation areas can be 2D or 3D spaces; after sampling the observation areas, firstly conducting compressed sampling, then conducting background updating through two moving averages, adopting the algorithm with anti-noise ability to obtain a compressed sampling value of background-subtracted images; detecting interested targets through threshold test by utilizing the compressed sampling value of the obtained background-subtracted images; and performing the analytical algorithm on the interested targets. The compressed sensing-based multi-view target tracking and 3D target reconstruction system and method is mainly applied to the design and manufacture of an image sensor.

Description

technical field [0001] The invention belongs to the field of image processing, and in particular relates to a system and method for multi-view object tracking and 3D object reconstruction based on compressed sensing. Background technique [0002] With the advancement of camera technology, the network of cameras widely used in surveillance systems, medicine, smart rooms, etc. continues to develop. The information obtained by multiple cameras can be used to locate the target, so as to track the occluded target or reconstruct the 3D target. In this context, it is particularly important to design a distributed algorithm that is robust to background changes in the observation area, has low communication overhead, and has a variable number of cameras. [0003] In the traditional method, a large amount of data with complete structure but high redundancy is captured by the camera, and then, if the smart camera can perform local processing, the acquired image will be preprocessed be...

AI Technical Summary

Problems solved by technology

[0005] The present invention aims to overcome the deficiencies of the prior art, and avoid the traditional multi-view video analysis system with large transmission data bandwidth, complex camera end and high power consumption, algorithm complexity without adaptability, weak noise resistance, and problems in practical applications. The disadvantage of poor scalability not only reduces the complexity of the coding end, reduces the transmission data bandwidth, but also improves the algorithm adaptability and signal noise resistance, and is robust to background changes. In order to achieve the above goals, the present invention adopts The technical solution is a system and method based on compressed sensing multi-view target tracking and 3D target reconstruction, including the following steps:

Images