光电跟踪测量设备电视系统受到自然条件的影响，往往捕获不到清晰的画面，很容易出现对比度极低的图像。针对此种情况，提出了一种用邻域差值来确定分段点的灰度变换法。通过计算像素5×5邻域的灰度差值，取最小的12个值之和ND。在大量实验的基础上确定合适的阈值T，与ND相比较，确定出图像边缘像素的灰度值范围，据此确定分段灰度变换公式的分段点。实验结果表明，在原图像对比度为1.19%的基础上，通过该算法处理后图像的对比度是原图像的18.79倍，是直方图均衡化后图像的6.97倍，是直方图规定化后图像的41.41倍，图像对比度达到了22.36%，图像的直方图也由单峰突出变为比较均衡的状态。该算法无论在数据上还是在视觉上都取得了很好的效果，满足了电视系统的跟踪要求，已经应用在光电跟踪测量设备上，有较好的实用价值。

The TV system of opto-electric tracking and measuring device, affected by natural conditions, often cannot capture focused images. It is prone to extremely low contrast image. This paper presents a gray-scale transform method with the difference of neighbor pixel to determine the segment points. First, by calculating the difference between the gray scale pixel 5×5 neighborhoods, ND is taken as the sum of the minimum 12 values. Then, on the basis of a large number of experiments, the appropriate threshold T is determined, with which the ND is compared to determine the edge pixel in the image range. Finally, according to the scope, the segment points of segmentation gray-scale transform formula are determined. The results show that after processed by this algorithm, the contrast is 18.79 times of the original image which is 1.19%, 6.97 times of the histogram equalization and 41.41 times of the histogram specification. The contrast has reached 22.36%, and the image histogram is also highlighted by the single peak to a more balanced state. The algorithm has obtained good data and visual effects, and met the requirements of TV system, which has been used in optical tracking device. The algorithm has good practical value.