An adaptive image thresholding algorithm by mean of background estimation in Gaussian scale space is proposed for thresholding images with uneven illumination. Firstly, a Gaussian scale space, which is produced by the convolution of a two-dimensional Gaussian function with an input image, is used to estimate the background image. After background subtraction, the objective image can be easily obtained to eliminate interference of uneven illumination. Secondly,γ correction is employed to enhance the image to highlight those darker ob jects. Finally, the thresholding result is extracted easily using the global valley-emphasis Otsu method. To test the effectiveness of the introduced scheme, image segmentation tests are carried out for document and non-document images with uneven illumination, and then comparisons on misclassification error (ME) and time expenditure are performed among the proposed approach, the biased field-based fuzzy c-means (FCM) method, the adaptive gray wave transformation thresholding scheme and the adaptive minimum error thresholding algorithm. The results show that the introduced method yields better visual quality and lower ME values than these three approaches.%为有效分割非均匀光照图像，提出一种在高斯尺度空间下估计背景的自适应阈值分割算法。首先，利用二维高斯函数对待处理图像进行卷积操作来构建一个高斯尺度空间，在此空间下进行背景估计，并采用背景差法来消除非均匀光照干扰，从而提取出目标图像；然后，采用γ矫正进行增强处理以突出较暗目标信息；最后，经强调谷底的最大类间方差法进行全局分割得到最终结果。为验证算法的有效性，对非均匀光照条件下文本图像以及非文本图像进行了测试，并与基于偏移场的模糊C均值方法、灰度波动变换自适应阈值分割算法和自适应最小误差阈值分割算法，在错误分割率和运行时间上进行了对比。实验结果表明，对比以上三种方法，该算法的分割结果更为理想。
展开▼
