Abstract:

A novel approach is proposed to detect moving dim infrared targets against different types of complex background. Background predictor is selected automatically based on image information. Meanwhile, a concept of “differentiation degree of local small targets” (DDLST) is proposed for the discrepancy of characteristics between the target and the background in complex background conditions, and an effective method is defined to quantitize it.

A new threshold function based on the DDLST is constructed by Monte Carlo methods to accomplish the target detecting algorithm in a single frame. Then, variable weighted pipeline filters are applied to extract the trace of moving dim targets. The experimental results demonstrate the good performance of this approach.

Key words: infrared image, background prediction, detection of moving dim targets, differentiation degree of small local targets, Monte Carlo methods