A heterogeneous remote sensing image registration method and device based on a multi-scale generative adversarial network

Abstract

The invention discloses a heterogeneous remote sensing image registration method and device based on a multi-scale generative confrontation network based on image segmentation. According to two sets of heterogeneous remote sensing image data, training samples are obtained; image segmentation is performed on input images to obtain segmented images; Construct a multi-scale generative adversarial network based on image segmentation; construct a generative adversarial network for image blocks; train two generative adversarial networks; perform homologous image registration between the generated image and the image to be registered. The present invention can constrain the spatial information consistency of images generated by the generative adversarial network, and provide an image conversion solution with full size, high resolution, and consistent spatial information, and simplify the registration problem of heterogeneous images into the registration problem of homogeneous images. Construct a unified heterogeneous remote sensing image registration framework.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a heterogeneous remote sensing image registration method and device based on a multi-scale generative adversarial network for image segmentation. Background technique [0002] Heterogeneous remote sensing image registration is a technology to register the geometric information of data acquired by different sensors in the same area. It is a key technology for the automatic processing of massive remote sensing data. Its calculation accuracy will directly affect the fusion and mosaic of heterogeneous remote sensing images. The precision of the operation to be applied subsequently. However, due to differences in imaging mechanism, shooting time, satellite attitude, and sensor parameters, there are often large differences in size, rotation, and translation between the data acquired by sensors of different nature. At the same time, there are often non-linear differenc...

AI Technical Summary

Problems solved by technology

However, the current image conversion methods based on machine learning are difficult to achieve high-quality image generation for high-resolution remote sensing data, making it difficult to apply machine learning-based heterogeneous remote sensing image registration methods to high-resolution heterogeneous remote sensing data

Images