Semi-supervised multispectral and panchromatic remote sensing image fusion method and system

Abstract

The invention discloses a semi-supervised multispectral and panchromatic remote sensing image fusion method and system. According to the scheme, a V channel of a multispectral image of an HSV color space is extracted, and the V channel and a panchromatic image are input into a twin network in a generator for encoding; then the features obtained by encoding are spliced and input into a decoder, andimage reconstruction is carried out after layer-skipping connection is carried out on the features obtained by convolution of each layer in the encoder and the corresponding layer of the decoder, sothat a fused V channel is obtained; the fused V channel is identified by using an identifier and the V channels of the panchromatic image and the multispectral image respectively, and an identification result is fed back to a generator to adjust generator parameters until the identifier passes identification;H, S and the fused V channel are spliced to obtain an HSV fused image, and the HSV fused image is converted into an RGB image. According to the method, label images do not need to be obtained, the calculation method is simple, and various indexes of the fused images have great advantages compared with existing methods.

Description

technical field [0001] The invention relates to the field of image processing (fusion), in particular to a semi-supervised multispectral and panchromatic remote sensing image fusion method and system. Background technique [0002] Due to the limitations of imaging sensor storage and signal transmission bandwidth, most earth observation satellites can only provide panchromatic (PAN) images with low spectral resolution and high spatial resolution and multispectral (MS) images with high spectral resolution and low spatial resolution. )image. Because the former is a single-band image, this type of image cannot obtain the color of the ground object, but has high spatial resolution; while the latter obtains an image with multi-band spectral information because the sensor acquires multiple bands of ground object radiation. Different bands are assigned different RGB values ​​to obtain low-resolution color images. Remote sensing images with high spatial resolution and high spectral...

AI Technical Summary

Problems solved by technology

Although the method based on multi-resolution analysis has high spectral fidelity, it usually cannot provide enough spatial information; the method based on sparse learning can use the dictionary to obtain high-resolution multi-spectral images, and the spectral distortion is small

However, most of the existing methods based on sparse learning have complex models, high computational complexity, and cannot preserve spectral information well

In addition, the need for label images (remote sensing images with both high spatial resolution and hyperspectral resolution) is a common problem of this type of method, but in practical applications, the acquisition of label images is a big problem

Images