thresholding an image with bright zones

Question

I am developing an app for iOS with openCV that take a picture from a monitor and extract a curve, but when the image has some bright zones after thresholding, I don't get the complete curve but some black zones

Original image

processed image after thresholding

        original = [MAOpenCV cvMatGrayFromUIImage:_sourceImage];   
        cv::threshold(original, original, 70, 255, CV_THRESH_BINARY | CV_THRESH_OTSU);

        findContours(original, contours, hierarchy,CV_RETR_CCOMP, CV_CHAIN_APPROX_SIMPLE );

        cv::Mat skel(original.size(), CV_8UC1, cv::Scalar(0));

        int idx = 0;
        for(; idx >= 0; idx = hierarchy[idx][0])
        {
            if (contours[idx].size()>250 && idx>-1){
                cv::Scalar color( 255,255,255);
                drawContours(skel, contours, idx, color, CV_FILLED, 8, hierarchy);
            }
        }
        cv::threshold(skel, skel, 100, 255, CV_THRESH_BINARY_INV);
        cv::erode(skel,skel,cv::Mat(),cv::Point(-1,-1),2);

So how I can process the image to extract the curve when the image have some bright zones like the example

FiReTiTi · Accepted Answer · 2016-06-27T08:56:24.333

0

When you have a background with an uneven illumination, you may want to apply first a White Top-Hat (or here for MatLab, and here for OpenCV).

Here is the result I got using a structuring element of type disk with radius 3.

Then, whatever thresholding method you choose will work.

edited Jun 27 '16 at 08:56

answered Jun 27 '16 at 07:20

FiReTiTi

5,597
12
30
58

I get a black image when I apply this code: cv::morphologyEx(imgColor, imgColor2, cv::MORPH_TOPHAT, cv::getStructuringElement(cv::MORPH_ELLIPSE,cv::Size(3,3))); can you give me an example, thank you!! – David Eduardo Ramirez Jun 27 '16 at 13:26
I don't use OpenCV, but maybe you are performing a black top-hat. A white top-hat is: Image - Opening(Image). Try this version. – FiReTiTi Jun 27 '16 at 18:12
It worked, but increasing the kernel size to > 16, thanks – David Eduardo Ramirez Jul 04 '16 at 18:21
using a RECT element make the process faster – David Eduardo Ramirez Jul 04 '16 at 21:57
It's normal because a rectangle can be decomposed into two segments, so you drop the complexity. 16 seems surprisingly big for this images, but maybe you have other images that require such by SE. – FiReTiTi Jul 05 '16 at 08:13

score 0 · Answer 2 · answered Jun 27 '16 at 07:46

0

Wouldn't be sufficient to use Otsu's thresholding?

Code fragment:

import cv2

image = cv2.imread('d:/so.jpg', cv2.IMREAD_GRAYSCALE)

threshold, thresholded = cv2.threshold(image, 0, 255, type=cv2.THRESH_BINARY + cv2.THRESH_OTSU)
print(threshold)

cv2.imshow('so', image);
cv2.imshow('thresholded', thresholded)
cv2.waitKey(0)

answered Jun 27 '16 at 07:46

Pavel Mach

36
4

I already use Otsu, but how you see the curve is still noisy in the bright spots – David Eduardo Ramirez Jun 27 '16 at 13:29

thresholding an image with bright zones

2 Answers2