High-resolution remote sensing image building change detection system under single-class classification framework

Abstract

The invention discloses a high-resolution remote sensing image building change detection system under a single-class classification framework, and relates to the technical field of building change detection. The system comprises a feature extraction module, a single-class information extraction module and a change correction module. The feature extraction module is connected with a database and is used for acquiring two original images and performing morphological building index feature extraction on the original images; the single-classification information extraction module is used for extracting morphological building index features and spectral features of high-resolution remote sensing images, performing multi-feature fusion on the morphological building index features and the spectral features, and obtaining an object-level building change detection result from an object-oriented angle by using a positive sample single classifier; the change correction module is used for correcting the change detection result of the object-level building obtained by the single-classification information extraction module; the system has certain robustness, and compared with an existing building change detection algorithm, the system has a better detection effect.

Description

technical field [0001] The invention relates to the technical field of building change detection, in particular to a high-resolution remote sensing image building change detection system under a single-class classification framework. Background technique [0002] Buildings are one of the basic spatial structures of a city, and their spatial change distribution is of great significance to urban planning and monitoring of illegal construction. Traditional building change detection requires a lot of manual field surveys, which is time-consuming and laborious, and it is difficult to quickly obtain large-scale building changes. Building change detection based on high-resolution remote sensing images can quickly identify its accurate large-scale spatial distribution changes, saving a lot of manpower and material resources. Therefore, it is of great research significance to study building change detection based on high-resolution remote sensing images; [0003] At present, with th...

AI Technical Summary

Problems solved by technology

After being improved, a single classifier is used to train and predict only one class of positive samples, and it is successfully applied to the land change survey of Landsat8 medium-resolution remote sensing images, and good experimental results have been achieved. Negative samples participate in the multi-class classification problem of training at the same time, but its

Images