图像中的噪声会直接影响图像分割质量,为快速、准确地识别含噪图像中的目标,提出一种基于直方图预处理与BF算法的含噪图像分割方法。该方法通过小波变换抑制图像中的噪声,分析增强图像的直方图特点以缩小分割阈值的分布范围,以二维最大类间方差为原则设计分割目标函数,利用BF算法快速搜索最优分割阈值。实验结果表明,该方法在收敛速度、稳定性和分割效果三个方面均优于基于遗传算法、人工鱼群算法等其他群体智能的分割方法。
Image noise may have a direct influence on the quality of image segmentation. In order to distinguish targets from noise-polluted image quickly and accurately, this paper proposes a method based on histogram preprocessing and BF algorithm for noisy image segmentation. In this method, discrete wavelet transform is used to suppress the noise in the image firstly. Secondly, the histogram feature of the denoised image is analyzed to shrink the distribution range of the optimal threshold. Then, two-dimensional Otsu is selected as the segmentation objective function, and bacterial foraging algorithm is employed to find the optimal threshold in parallel. Experimental results show that this method performs better than some other methods based on swarm intelligence like genetic algorithm, artificial fish swarm algorithm as far as convergence speed, stability and segmentation effect are concerned.