摘要
提出了基于改进的脉冲耦合神经网络(PCNN)与Otsu的图像增强新方法。该方法对 PCNN进行了改进,而用改进后的PCNN进行图像去噪处理,继而用Otsu方法寻找最佳灰度阈值后进行图像增强。仿真实验表明,该方法滤波后信噪比(PSNR)为18.930 5,而高斯滤波为 5.408 7;同时又能根据图像灰度性质自动选取最佳阈值,并对自适应分割后图像进行不同的灰度变换,使图像得到有效增强。仿真结果证明了该方法的有效及合理性。
A new enhancement method of image based on pulse coupled neural network (PCNN) and Otsu was proposed, in which improved PCNN is used to remove the noise and Otsu is used to search best gray threshold value for image enhancement. The results of experiments show the algorithm can remove noises more effectively than traditional method (PSNR is 18.9305 for the proposed algorithm, 5.4087 for gauss algorithm), and the best threshold can also be achieved. In additional, different gray switch function is taken on the basis of the best threshold. Consequently, the remarkable effect of image enhancement is gained.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2005年第3期358-362,共5页
Journal of Optoelectronics·Laser
关键词
脉冲耦合神经网络
PCNN
Otsu理论
阈值
图像增强
Adaptive algorithms
Computer simulation
Image segmentation
Interference suppression
Neural networks
