An Extraction of Infrared Occluded-Object Based on Maximum Variance and Negative Selection

Object extraction is an aspect of image segmentation. When the edge is clear, the object will be easily extracted, while the edge is fuzzy, the extraction will become hard. As an important method of the immune recognition, negative selection algorithm is a simulation of the immune cellspsila maturation, and its detection mechanism is especially suitable for the extraction of a specific target in the image. Based on the infrared image features and the maximum variance of the temperature threshold function method, this paper puts forward new detector-generation rules and creates an immune negative selection algorithm suitable to the image object detection. It receives a relatively ideal detective result, applying the algorithm to the extraction of a infrared occluded-object.