Ground object classifying method and device

Abstract

The invention relates to the technical field of remote sensing, and provides a ground object classifying method and device. The ground object classifying method includes the steps that a plurality of attribute section characteristics of a hyperspectral image are extracted to obtaina first image; attribute section characteristics of a laser scanning image are extracted to obtain a second image; the first image and the second image are fused to obtain a third image; a preset convolutional neural network is adopted to conduct characteristic extraction and classification on the third image to obtain ground object classifying results. According to the ground object classifying method and device, abundant spectral information of the hyperspectral image and accurate elevation information of the laser scanning image are complementary, and the problem is solved that due to inaccurate spectral information, the ground object classification is limited. Besides, the convolutional neural network is adopted to conduct the characteristic extraction and the classification, the requirements on the number of training samples are reduced, and at the same time, the accuracy of the ground object classification is improved.

Description

technical field [0001] The invention relates to the field of remote sensing technology, in particular to a method and device for classifying ground objects. Background technique [0002] Hyperspectral imagery is a cutting-edge technology in the field of remote sensing, which can acquire hundreds of spectrally continuous bands. Compared with panchromatic and multi-spectral remote sensing images, hyperspectral images have higher spectral resolution and can provide richer object information, thereby better identifying object objects. However, hyperspectral images cannot well solve the problems of building shadows and cloud coverage in complex urban areas. In addition, when classifying objects in more complex urban areas, hyperspectral images cannot effectively distinguish different objects composed of the same material. Ground objects, therefore, need to extract more separable spatial spectral features. [0003] Spatial structure features based on morphological attribute prof...

AI Technical Summary

Problems solved by technology

However, hyperspectral images cannot well solve the problems of building shadows and cloud coverage in complex urban areas. In addition, when classifying objects in more complex urban areas, hyperspectral images cannot effectively distinguish different objects composed of the same material. Therefore, it is necessary to extract more separable spatial spectral features

[0003] Spatial structure features based on morphological attribute profiles can effectively extract multi-scale structural information in hyperspectral images, but due to the complexity and diversity of hyperspectral images, a single feature can only describe hyperspectral images. In the identification and classification of , it is difficult to obtain enough training samples, and the calculation cost is high

Images