Landslide identification method based on combination of symmetric deep network and multi-scale pooling

Abstract

The invention discloses a landslide identification method based on combination of a symmetric deep network and multi-scale pooling, and the method comprises the steps: constructing a training set image pair and a test set image pair through employing a large-size high-resolution dual-temporal remote sensing image; giving preprocessing algorithm parameters and network operation parameters; generating a difference image corresponding to the training set by using the training set image pair, and generating a difference image corresponding to the test set by using the test set image pair; carryingout multivariable morphological reconstruction on the difference image to remove noise and non-landslide areas; preprocessed images are input into the deep network MP-designed by the invention; in the SDNN, carrying out network model training until the network converges; input of pre-processed test images to a network MP- In the SDNN, a landslide identification result is output, the problems thatan existing landslide identification method depends on a traditional feature descriptor, threshold selection is carried out, and a landslide area is difficult to accurately identify due to noise sensitivity can be solved, the identification precision of a real landslide area is improved, and the method has the advantages of being high in identification precision, rapid and effective.

Description

technical field [0001] The invention belongs to the field of image processing technology and pattern recognition, in particular to a landslide recognition method based on a symmetrical deep network combined with multi-scale pooling. Background technique [0002] Landslides are a common natural disaster triggered by factors such as seismic activity, heavy rainfall, hillside construction, and human activities. In recent years, due to the increasing probability of sudden landslides year by year, and with the rapid development of remote sensing technology, the introduction of high-resolution technology makes remote sensing landslide images have richer ground object information, and the shape, texture, and context information of landslides are more complex. , so the study of high-precision landslide identification methods has become a hot topic to meet the needs of fast and effective landslide identification methods in practical applications. [0003] Early landslide identificat...

AI Technical Summary

Problems solved by technology

[0006] In order to overcome the shortcomings of the above-mentioned prior art, the purpose of the present invention is to provide a landslide identification method based on a symmetric deep network combined with multi-scale pooling, which can solve the current Some landslide identification methods rely on traditional feature descriptors, threshold selection, and sensitivity to noise, which makes it difficult to accurately identify landslide areas, etc., to improve the identification accuracy of real landslide areas, and has the characteristics of high identification accuracy, fast and effective

Images