Reflection reduction in image

Question

Lighting problems are common and difficult. How does one detect and reduce light reflection to save more information from an image? I have tried several methods with OpenCV and Python without luck.

(Image with reflection)

(Image without reflection)

I tried converting to HSV color space, and apply Histogram Equalization to the V channel, with Clahe equalization:

import cv2 import numpy as np  image = cv2.imread('glare.png')  hsv_image = cv2.cvtColor(image, cv2.COLOR_RGB2HSV) h, s, v = cv2.split(hsv_image)  clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8,8)) v = clahe.apply(v)  hsv_image = cv2.merge([h, s, v]) hsv_image = cv2.cvtColor(hsv_image, cv2.COLOR_HSV2RGB)  cv2.imwrite('clahe_h.png', hsv_image) 

results:

As well I tried thresholding to find bright pixels and than use Image Inpainting to replace them with neighbouring pixels.

import cv2 import numpy as np  image = cv2.imread('glare.png')  gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) blurred = cv2.GaussianBlur(gray, (3,3), 0) thresh = cv2.threshold(blurred, 225, 255, cv2.THRESH_BINARY)[1]  dst_TELEA = cv2.inpaint(image,thresh,3,cv2.INPAINT_TELEA) cv2.imwrite('after_INPAINT.png',dst_TELEA) 

results: (after threshold)

Answer 1

There's no general method for effective glare removal.

HSV + CLAHE is a good and common start, but industry methods "cheat" by assuming some information about the subject (human face, fruit on a conveyor belt, retina through an ophthalmoscope), and sometimes information about the lighting (white ceiling light with very sharp edges, for the images in this question).