A Review of Camouflaged Target Detection Research

Camouflaged targets are highly integrated with the background, with blurred features and difficult to detect. This paper reviews the research on conventional algorithms and convolutional neural networks and deep learning based algorithms on camouflaged targets, and analyzes the advantages and disadvantages of different algorithms. Then the current data sets of camouflaged people made public by relevant scholars are introduced, the problem of high leakage rate of conventional target detection algorithms on camouflaged small-scale targets is pointed out, and the evaluation indexes of target detection algorithm performance are summarized. Finally, the directions that should be researched on camouflaged target detection in the future are prospected.