OpenCV预处理图像，以获得更好的阈值处理结果

Question

在OpenCV中，如何预处理图像以获得更好的阈值结果？我的输入图片是

base image

我正在使用各种阈值方法，如：

修改

应用CLAHE (Contrast Limited Adaptive Histogram Equalization)会显着降低噪音量，但提取的文字仍然模糊不清。有没有办法让文字更加清晰＆＃34;？

clahe = cv2.createCLAHE(clipLimit=1.0, tileGridSize=(3,3))
inputImage = clahe.apply(inputImage)

END OF EDIT

invertRet, invertColors = cv2.threshold(inputImage,127,255,cv2.THRESH_BINARY_INV)

gaussianThreshold = cv2.adaptiveThreshold(inputImage, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV, 11, 2)

adaptiveMean = cv2.adaptiveThreshold(inputImage, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY_INV, 11, 2)

otsuRet, otsu = cv2.threshold(inputImage,0,255,cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)

但它们都显示出很多噪音：

gaussian thresholding