Visual capture detection method based on convolutional neural network

The invention discloses a visual capture detection method based on a convolutional neural network. The method comprises the following steps: obtaining an RGB picture and a depth picture of an object to be captured; the method comprises the following steps: preprocessing an input RGB picture and a depth picture to obtain a three-dimensional space coordinate of each pixel point of the RGB picture; removing the background of the RGB picture; feature extraction and feature fusion are carried out on the RGB picture with the background removed; performing target detection and grabbing detection operation on the fused features based on a target detection network and a grabbing detection network respectively at the same time to obtain target detection features and grabbing detection features; performing alignment operation on the features after target detection and the grabbing detection features to obtain the joint relevance of a target detection result and a grabbing detection result; and screening out a final grab