Pixel-by-pixel classification method, storage medium and classification equipment

Abstract

The invention discloses a pixel-by-pixel classification method, a storage medium and classification equipment. The method comprises the steps: reading a multispectral image from a data set, and the multispectral image comprises registered PAN image data and MS image data, and a corresponding class label ground truth image; fusing the common features to obtain MSHPAN image data; determining a training set and a test set; preprocessing, and designing a progressive fusion network based on the adaptive receptive field network module A and a B module extracted from spectral information of an adaptive channel; and training to obtain a classification model, and classifying the test set to obtain the category of each pixel point in the test data set. According to the invention, spatial informationand spectral segment information are extracted in a self-adaptive manner, and then are fused gradually.

Description

technical field [0001] The invention belongs to the technical field of computer vision image processing, and specifically relates to a pixel-by-pixel classification method, storage medium and classification equipment based on progressive fusion of adaptive receptive fields and adaptive channels, which can be used for remote sensing such as environmental monitoring, land cover, and urban construction. In related fields of image classification. Background technique [0002] In recent years, with the support of advanced equipment and technology, many earth observation satellites can acquire panchromatic images (PAN) with relatively high spatial resolution and multispectral images (MS) with rich spectral information within the same coverage area. Therefore, this feature complementarity between PAN data and MS data provides an important development potential for fusion classification in the field of remote sensing. [0003] Multi-resolution data fusion with complementary feature...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is to provide a pixel-by-pixel classification method, storage medium and classification equipment based on the progressive fusion of adaptive receptive fields and adaptive channels, through adaptive space and automatic A progressive fusion network adapted to spectrum segment selection, which solves the problems of simple and crude fusion methods and low classification accuracy in existing technologies

Images