图像中标定符号的定位与识别是进一步计算数字图像对应的空间距离的依据。本文应用Hough变换对图像标定符号进行定位,克服了强噪声背景下单独采用灰度特征定位的缺陷。定位后对图像进行分割,并与传统方法比较。实验证明,本方法对在图像中所占比例较小的标定符号有较好的定位效果,并结合标定符号的灰度特征对强噪声背景下的标定符号进行准确分割。
The localization and recognition of calibration signs is the basis to further calculate the spatial distance in digital images. This paper proposes a novel method of calibration sign localization based on Hough transform, which overcomes the deficiency of localization based on gray scale character solely. The novel method is compared with traditional methods after segmenting the image. It is proved by experiments that the method presented in this paper is more effective in the localization of calibration signs that has a less proportion of the whole image, furthermore, combined with gray scale character, could segment calibration signs from strong noisy background accurately.