High resolution ratio remote-sensing image division and classification and variety detection integration method

Abstract

The invention discloses an integrated method for segmentation, classification and change detection of high-resolution remote sensing images based on multi-level set evolution. Its characteristics are: (1) Image preprocessing (radiation, registration and feature filtering); (2) Establishing segmentation and classification models of multi-level set evolution, using GIS data after registration to automatically determine the initial contour of each level set function, Segment and classify the first phase image; (3) still adopt the model described in (2), optimize the initial profile of each level set function, and use incremental segmentation and classification for the second to T phase images; (4) Taking the segmented object as a unit, compare the classification results of the i-th and i+1-th adjacent time-phase images to determine the change area; (5) return to (3) until the completion of all T time-phase images Segmentation and classification and change detection. Advantages: Compared with the traditional pixel-oriented K-means method, the accuracy of classification and detection has been improved, and it is suitable for change detection of sequence remote sensing images, and has wide applications in disaster monitoring and land resource investigation.

Description

technical field [0001] The invention belongs to the technical field of computer remote sensing image information extraction, and relates to an integrated method for segmentation, classification and change detection of high spatial resolution remote sensing images based on multi-level set surface evolution. Background technique [0002] Remote sensing image information extraction is a basic and important research content in the fields of computer and remote sensing. However, with the rapid development of sensor technology, satellite communication technology, and computer technology, remote sensing images are increasingly showing the characteristics of "three multiples", that is, multiple sensors, multiple spatial or spectral resolutions, and multiple temporal phases, which make remote sensing images automatically Information extraction is more difficult. [0003] In recent years, high spatial resolution (spatial resolution less than or equal to 5 meters, hereinafter referred...

AI Technical Summary

Problems solved by technology

In addition, in the literature Dell'Acqua, F.Gamba, PPrevedini, P., Level-set based extraction and tracking of meteorological objects in satellite images, IGARSS 2000.vol.2, page(s): 627-629, using level-set The dynamic evolution characteristics of the theory designed a cloud target tracking system for low- and medium-resolution meteorological satellite data. The above work confirmed the applicability of the level set theory in single-target change detection in many time-phase remote sensing images, but for land use surveys For the change monitoring of multiple types of targets involved in the task, the above method is invalid, and a multi-level set (the number of level set functions is greater than 1) segmentation or classification method is required.

Images