Sonar Images De-noising Based on Pulse Coupled Neural Networks

In order to reduce the noise effect in sonar images, the algorithm combined morphology with median filter was proposed to wipe off the Gaussian and pulse noise simultaneously which was based on the simplified pulse coupled neural networks (PCNN). Using the PCNN’s output images we could confirm the pixel’s position corresponding to noise, and keep the internal integrity of targets by the morphological filter. Then the median filter was used for suppressing the pulse noise, and the adjustment of the gray-level values in iterative PCNN output images would reduce the Gaussian noise. Simulation results show that our method’s effectiveness is better than general noise suppression algorithms and has small time consumption, so it is more useful and fits for real-time sonar operating.