Abstract:

Focusing on the problem of ship detection in poor imaging conditions,which means
here,being disturbed by lots of ragged clouds in optical images,a robust ship detection
algorithm is proposed based on the characteristics of the background and the target itself.
Firstly,it extracts ROI using two parameters and local CFAR. Then it extracts the wake
information by the improved local Radon transform and a peak value extraction algorithm in
succession.The experiments of satellite images show that this method can reduce false alarms
caused by ragged clouds effectively .

Key words: ship detection;ragged clouds;CFAR;local Radon transform;wake