Study Of Infrared Small And Weak Target Detection In Complex Background

The detection of infrared small targets under complex background is one of thehottest topics in the field of image processing, and the exploration of new small targetdetection algorithm has great significance, as well as the study of how to apply theexisting algorithm to the detection of small targets correctly.After the description of the characteristics of the infrared image, we introduce severalcommon image preprocessing algorithm and target detection technology, then we sum uptheir performance and analyse their applicable conditions. They mainly includebackground suppression algorithm and threshold filtering algorithm. Based on thestudying and comparing of these algorithms, a new infrared small target algorithm wasproposed in this paper: use the connected domain filtering algorithm based on the layeredthreshold method to suppress the background, and use pipeline filter based on connecteddomain matching algorithm to filter the false targets. The connected domain filteringalgorithm based on multi-threshold layering, delaminate the picture by using theindependence between the target and the background and the highly relativity of the pixelin the target. Then we use its area and shape and gradient character to distinguish thetarget domain out of the background. The gradient means the gradient between theconnected domains. By adjusting the target area threshold, this algorithm can be adaptedto pictures that include all kinds of targets. Having came up with this method, we discussthe noise filtering algorithm and the connected domain algorithm calculation method thatassociate with this algorithm and optimize it to make it support real-time in the lastchapter of this paper. In the end, the experiment shows that this method achieve thebalance of the time performance and the accuracy of detection, and it has a good detectionperformance.