基于多帧移位叠加的红外小目标检测方法

Dim target detection algorithm based on multi-frame indexing accumulation

由于基于能量积累的各种算法对预检测的目标漏检和噪声密度都比较敏感 ,所以要求预检测在注重目标检出率的同时应尽量减小虚警。例如当噪声密度大于 3%时高阶相关法检测性能大幅下降 ,Hough变换法在噪声密度较大时不但性能下降 ,而且运算量也增加很多。基于此 。

The probability of the target detection must be increased and the intensity of the noise must be decreased in pre\|detection because the algorithms based on energy accumulation are sensitive to the misdetection of the target and the intensity of the noise. When the intensity of the noise is larger than three percent, for example, the capability of the detection using high order correlation will decrease rapidly. And when Hough transform is used in target detection in the condition of large noise intensity, not only the calculation would be very heavy but also the capability of detection would be not ideal. Therefore, a self\|adaptive threshold pre\|detection method using multi\|frame indexing accumulation based on energy accumulation is developed.