Spatial background modeling method based on time domain information characteristics

The invention aims to provide a space background modeling method based on time domain information characteristics. The space background modeling method comprises the following steps: A, reading multiple frames of images to be detected; b, constructing spatial neighborhood gradient filtering function...

Full description

Saved in:
Bibliographic Details
Main Authors CHEN HUAJIN, WU ANQING, LI JULIU, FAN XIANGSUO, YAN CHUAN, YAO JIALE, WANG NAYI
Format Patent
LanguageChinese
English
Published 12.03.2021
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:The invention aims to provide a space background modeling method based on time domain information characteristics. The space background modeling method comprises the following steps: A, reading multiple frames of images to be detected; b, constructing spatial neighborhood gradient filtering functions in multiple directions, and calculating and obtaining gradient filtering function values of each pixel point in multiple directions; c, filtering the multi-frame number of images to be detected to obtain a multi-frame number of filtered images; d, constructing a training sample matrix for to-be-detected images of multiple frames, and then performing decentralization processing, covariance calculation and singular decomposition on the training sample matrix to construct a feature space matrix;and performing matrix reconstruction by using the feature space matrix to obtain background modeling images corresponding to the to-be-detected images of each frame number. The method overcomes the defects in the prior art, a
Bibliography:Application Number: CN202011468281