Method for detecting open-pit mine field in remote sensing image based on deep learning

Abstract

The invention relates to a method for detecting an open-pit mine field in a remote sensing image based on deep learning, and belongs to the field of target detection. Based on a Mask R-CNN network, aregion generation network MA-RPN based on mixed attention and a feature pyramid network ET-FPN based on expansion are designed on the network, a mixed attention mechanism is introduced into the MA-RPN, an attention module is added into the region generation network, and key features of an open-pit mine field in a picture are identified through an attention mask. Therefore, the model is helped to learn an open-pit mine field area which needs to be concerned; the ET-FPN is provided with a pyramid layer specially used for open-pit mine field detection and used for extracting information of an open-pit mine field in a remote sensing image, and the expanded feature pyramid layer is fed back to a subsequent detector for further positioning and classification. The method can achieve the detectionof the open-pit mine field in the remote sensing image.

Description

technical field [0001] The invention belongs to the field of target detection, and relates to an open-pit mine detection method in remote sensing images based on deep learning. Background technique [0002] The traditional object detection method can only have better results in specific scenes, and it is difficult to achieve better results when the lighting and other environments change, and it is prone to missed and false detections; in addition, the accuracy of traditional methods is not high High, it is difficult to apply to the detection of open-pit mines in remote sensing images. The present invention has invented a method specifically for the detection of open-pit mines in remote sensing images. Based on the Mask RCNN network, a hybrid attention-based The region generation network (MA-RPN) based on the extended feature pyramid network (ET-FPN), the deep learning model of the present invention has a better effect on the detection of open-pit mines in remote sensing imag...

AI Technical Summary

Problems solved by technology

[0002] The traditional object detection method can only have better results in specific scenes, and it is difficult to achieve better results when the lighting and other environments change, and it is prone to missed and false detections; in addition, the accuracy of traditional methods is not high High, it is difficult to apply to the detection of open-pit mines in remote sensing images. The present invention has invented a method specifically for the detection of open-pit mines in remote sensing images. Based on the Mask RCNN network, a hybrid attention-based The regional generation network (MA-RPN) and the feature pyramid network (ET-FPN) based on the extension, the deep learning model of the present invention has a better effect on the detection of open-pit mines in remote sensing images

Images