一种对光照具有鲁棒性的图像分割方法*

研究了灰度图像的OTSU(最大类间方差)自动阈值分割法。OTSU方法作为一种单一阈值的分割方法，当图像受光照和反射光等的影响明显时，将会出现严重的误分割现象。考虑到OTSU方法的最大类间方差化的思想，根据灰度图像的像素点灰度的直方图分布、空间分布，提出了一个新的分割阈值方法。先根据OTSU方法的特点自设计一个函数，对图像进行变换，以便后面的处理，再对其图像以改进的OTSU方法进行分割。通过对化学实验中两种液体的拍摄图片及数字图像处理中标准图片进行试验，理论分析与实验结果表明：该方法能够对受光照及反射光影响大的图像实现正确的分割，将目标图像清晰地从背景中分割出来。

Robust Method of Image Segmentation for Image Affected by Illumination and Reflexion

Studies the method of OTSU automatic thresholding segmentation.As a method of single threshold segmentation,the effect is bad when the image is affected by illumin and reflexion. A new method of automatic threshold selection is proposed according to the gray level statistics about the pixel of the local target and background region. Firstly the pixel in image is transfered by a special function, then the image is devided into two parts by an improved method of OTSU automatic thresholding.Theoretical analysis and practical outcomes show that correct image segmentation can be obtained for an image badly affected by illumination and reflexion.