Dense feature matching method and device for binocular three-dimensional reconstruction, equipment and medium

The invention provides a dense feature matching method and device for binocular three-dimensional reconstruction, equipment and a medium, and the method comprises the steps: collecting and obtaining a binocular image which comprises a left image and a right image; extracting the binocular image by using a convolutional neural network to obtain a first feature map, wherein the first feature map comprises a first shallow-layer feature map and a first deep-layer feature map; performing feature matching and view angle transformation on the first deep feature map to obtain a view angle transformation map; extracting the view angle transformation graph by using a convolutional neural network to obtain a second feature graph, wherein the second feature graph comprises a second shallow-layer feature graph and a second deep-layer feature graph; fusing the first deep feature map and the second deep feature map to obtain a fused feature map; and carrying out feature matching on the fused feature map. According to the me