Micro operating space target detection algorithm based on improved Faster RCNN
Faster RCNN is introduced into the target detection of micro operating system.Aiming at the problems that there is scale change in target to be detected in micro operating space and scale of the target to be detected is too small and the feature is not obvious when microscope magnification is small,a micro operating space target detection algorithm based on improved Faster RCNN is proposed.Depth residual network with superior performance in image classification task is used to extract image features.Recursive feature pyramid network is introduced to fuse the features.The sampling strategy of the regional recommendation network is improved,and the loss function is optimized.Experimental result shows that the improved Faster RCNN algorithm can effectively solve the problems caused by the change of target scale and too small target scale.Compared with the general target detection algorithm,the algorithm has higher accuracy,faster speed and practical application value.
micro operating spacetarget detectionfeature extractionregion proposal network(RPN)sampling strategyloss function optimization
万方数据
维普
