根据图像背景的局部平稳性,提出了通过逼近背景局部相关特性来估计图像背景的方法,并且采用 RBF 神经网络来逼近图像背景的这种局部相关性.文中根据径向基函数神经网络的训练要求,用遗传算法来训练神经网络中的各个参数。实验结果表明,训练后的神经网络背景抑制算法能够很好地抑制图像中的背景,并同时增强了目标信号。 According to the local smoothness of the image background, a method of estimating the background of the image by approximating the local correlation of the background is proposed, and the RBF neural network is used to approximate this local correlation of the image background. According to the training requirements of radial basis function neural network , Using genetic algorithm to train various parameters in neural network. The experimental results show that the trained neural network background suppression algorithm can well suppress the background of the image and enhance the target signal at the same time.