Remote sensing image semantic segmentation method and device fusing edge detection

Abstract

The invention provides a remote sensing image semantic segmentation method and device fusing edge detection. The method comprises the steps of manufacturing a training sample according to a high-resolution remote sensing image; respectively training an HED edge detection model and an image semantic segmentation model; performing edge detection on the remote sensing image to be segmented by using the trained HED edge detection model to generate an edge probability graph, and performing post-processing on the edge probability graph; performing image segmentation on the remote sensing image to besegmented by using a trained image semantic segmentation model; and superposing the result of the edge detection post-processing and the image after image segmentation based on a geographic positionto obtain a semantic segmentation vector result of the remote sensing image to be segmented. According to the method, the edge detection model is combined, so that the edge of the segmentation objectis more accurate, the precision dependence on the semantic segmentation model is reduced, the complexity of the large-scale image semantic segmentation model is greatly reduced, and the model has expansibility.

Description

technical field [0001] The invention belongs to the field of surveying and mapping remote sensing / image processing, and in particular relates to a remote sensing image semantic segmentation method and device with fusion edge detection. Background technique [0002] With the increasing amount of high-resolution remote sensing satellite resources and data, the ground object information obtained by remote sensing is becoming more and more refined, and it also brings greater challenges to image interpretation technology. Remote sensing image segmentation is an automatic or semi-automatic remote sensing image interpretation method, which can mark remote sensing image pixels into objects with semantic concepts. It is currently more and more widely used in remote sensing image analysis. Common remote sensing commercial software such as ENVI, Erdas, and PCI have integrated remote sensing image segmentation algorithms, and there are also powerful professional remote sensing image seg...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the inaccurate edge segmentation problem of the existing remote sensing image depth semantic segmentation algorithm, the present invention proposes a remote sensing image depth semantic segmentation method and device that incorporates an edge detection algorithm, which significantly improves the edge accuracy of the segmented object

Images