Mask R-CNN和无人机遥感在黄河“四乱”问题调查中的应用

The application of Mask R-CNN and UAV remote sensing in the investigation of the“four disorders”in the Yellow River

黄河“四乱”问题排查和整治是保障黄河流域高质量发展的重要环节。为了提高“四乱”问题调查工作的效率和准确性,本文采用一种基于Mask R-CNN和无人机遥感的调查方法,可以有效区分“四乱”目标及其干扰物。首先采用无人机航摄系统获取目标区域的正射影像;然后构建属于“乱建”问题的养殖大棚样本库;最后采用Mask R-CNN模型进行对象检测。试验结果表明,采用方法对养殖大棚目标检测的准确率可以达0.93,且能大幅提高调查效率。

The investigation and rectification of the“four disorders”problems of the Yellow River is an important link to ensure the high-quality development of the Yellow River basin.In order to improve the efficiency and accuracy of the investigation of the“four disorders”problem,an investigation method based on Mask R-CNN and UAV remote sensing is adopted,which can effectively distinguish the“four disorders”targets and their interferences.Firstly,UAV is used to fly the target area to obtain orthophoto images,and then a sample library of breeding greenhouse belonging to the problem of“disorderly construction”is constructed.Finally,Mask R-CNN model is used for object detection.The experimental results show that the accuracy of the method for the detection of breeding greenhouse targets can reach 0.93 respectively,and the investigation efficiency can be greatly improved.