A Target Detection Algorithm Based on New Definition of Information Entropy

The exponential behavior is very efficient in image processing, when used in target detection in this paper, it is efficient too, and it also overcomes the shortage of logarithmic behavior. Simulation results show that the two different definitions of entropy have approximately the same effect on the uncertainty of targets, but compared with the logarithmic behavior of Shannon's entropy, this new definition of entropy avoids the undefined and zero values of information.