传统的图像分割算法在处理高噪声显微图像时, 由于背景复杂, 很难得到目标完整的区域轮廓。通过对不同图像分割算法的性能进行对比, 提出了一种改进的二维最大熵阈值遗传算法结合数学形态学除噪分割的方法。首先用改进的二维最大熵阈值算法结合遗传算法对高噪声显微图像进行粗分割, 除去图像中大量的背景噪声, 然后运用数学形态学进行细分割, 滤除剩余少量杂质和孔洞, 提取出目标轮廓。实验结果表明改进的方法较传统分割方法具有更强的抗噪声能力及更快的处理速度, 有效地实现了高噪声显微图像的除噪分割。

Due to the complexity of background, the traditional image segmentation algorithm is difficult to get a complete outline of the target when dealing with the high noise microscopic image. By comparing the performance of different image segmentation algorithms, an improved two-dimensional maximum entropy threshold genetic algorithm combined with mathematical morphology is proposed. Firstly, the improved 2D maximum entropy threshold genetic algorithm is used to do the rough segmentation to remove a large amount of background noise in the image, and then the mathematical morphology is used to do the fine segmentation to filter out a small amount of remained impurities and holes and extract target profile. The experimental results show that the improved method has stronger ability to resist noise compared with traditional segmentation method, and the processing speed is improved greatly, the segmentation of high noise micro-images is implemented effectively.