标定板图像二值化中阈值自动选取的一种方法

A method of automatic threshold selection in calibration plate image binarization

在机器视觉测量系统中,采集到的标定板图像往往存在光强不均的情况,只用一个阈值对整幅图像进行二值化往往达不到预处理要求。将图像分为几个区域,对每个区域结合Otsu方法和统计学理论,利用均值和方差自动选取阈值,进而对此区域进行二值化。与传统的方法相比较,该算法计算量小,公式简单,实现容易,提高了处理速度,实验结果达到了预期要求。

In the machine visionmeasurement system,the light intensity is often distributed unevenly in the collected calibration plate image.So only selecting one threshold for the entire image binarization can not achieve the purpose.The method divides the entire image to several regions.With the Otsu method and the statistics theory,the threshold in each region is selected by the average and the variance.Then the region can make image binarization using of the selected threshold.Compared to the traditional method,this arithmetic has small amount of calculation.The formula and the realization of the arithmetic is simple.And the processing speed is accelerated.Finally,the result achieves the purpose which is expected.