Multispectral image segmentation method and device, electronic equipment and storage medium

Abstract

The invention provides a multispectral image segmentation method and device, electronic equipment and a computer readable storage medium. The method comprises the following steps: performing image segmentation on a to-be-processed multispectral image through a first segmentation sub-model included in an image segmentation model to obtain a corresponding first segmentation feature image; on the basis of the multispectral image and the first segmentation feature image, carrying out feature expansion on the multispectral image through a first feature expansion sub-model included in the image segmentation model, and obtaining a corresponding feature expansion image; and performing image segmentation on the feature expansion image through a second segmentation sub-model included in the image segmentation model to obtain a first target segmentation image corresponding to the multispectral image. According to the invention, the segmentation precision of the multispectral image can be improved.

Description

technical field [0001] The present application relates to computer vision technology, and in particular to a multispectral image segmentation method, device, electronic equipment, and computer-readable storage medium. Background technique [0002] The classification of remote sensing object information is semantic segmentation, which is a pixel-level classification and is used in many aspects such as automatic driving, medical image diagnosis, and remote sensing object classification. Limited by complex backgrounds, such as edge noise, occluded lighting perspective, etc., semantic segmentation has low accuracy (such as average intersection ratio). And with the embedding and design of complex network structures, the semantic segmentation model tends to have a large amount of calculation, the reasoning speed is slowed down, it is difficult to explain, and requires a large amount of data fitting, which increases the labeling time and training cost. [0003] In related technolo...

AI Technical Summary

Problems solved by technology

Limited by complex backgrounds, such as edge noise, occluded lighting perspective, etc., semantic segmentation has low accuracy (such as average intersection ratio)

And with the embedding and design of complex network structures, the semantic segmentation model tends to have a large amount of calculation, the reasoning speed is slowed down, it is difficult to explain, and a large amount of data fitting is required, which increases the labeling time and training cost

[0003] In related technologies, for the segmentation method of multispectral images, the public datasets of remote sensing images used for semantic segmentation applications are relatively poor, and traditional segmentation methods generally use random forest or texture primitive forest methods to construct classifiers for semantic segmentation , this type of method needs to manually make copy features, poor robustness, and low segmentation accuracy of multispectral images

Images