I have some color photos and the illumination is not regular in the photos: one side of the image is brighter than the other side.
I would like to solve this problem by correcting the illumination. I think local contrast will help me but I don't know how :(
Would you please help me with a piece of code or a pipeline ?
Convert the RGB image to Lab color-space (e.g., any color-space with a luminance channel will work fine), then apply adaptive histogram equalization to the L channel. Finally convert the resulting Lab back to RGB.
What you want is OpenCV's CLAHE (Contrast Limited Adaptive Histogram Equalization) algorithm. However, as far as I know it is not documented. There is an example in python. You can read about CLAHE in Graphics Gems IV, pp474-485
Here is an example of CLAHE in action:
And here is the C++ that produced the above image, based on http://answers.opencv.org/question/12024/use-of-clahe/, but extended for color.
#include <opencv2/core.hpp>
#include <vector> // std::vector
int main(int argc, char** argv)
{
// READ RGB color image and convert it to Lab
cv::Mat bgr_image = cv::imread("image.png");
cv::Mat lab_image;
cv::cvtColor(bgr_image, lab_image, CV_BGR2Lab);
// Extract the L channel
std::vector<cv::Mat> lab_planes(3);
cv::split(lab_image, lab_planes); // now we have the L image in lab_planes[0]
// apply the CLAHE algorithm to the L channel
cv::Ptr<cv::CLAHE> clahe = cv::createCLAHE();
clahe->setClipLimit(4);
cv::Mat dst;
clahe->apply(lab_planes[0], dst);
// Merge the the color planes back into an Lab image
dst.copyTo(lab_planes[0]);
cv::merge(lab_planes, lab_image);
// convert back to RGB
cv::Mat image_clahe;
cv::cvtColor(lab_image, image_clahe, CV_Lab2BGR);
// display the results (you might also want to see lab_planes[0] before and after).
cv::imshow("image original", bgr_image);
cv::imshow("image CLAHE", image_clahe);
cv::waitKey();
}
The answer provided by Bull is the best I have come across so far. I have been using it to. Here is the python code for the same:
import cv2
#-----Reading the image-----------------------------------------------------
img = cv2.imread('Dog.jpg', 1)
cv2.imshow("img",img)
#-----Converting image to LAB Color model-----------------------------------
lab= cv2.cvtColor(img, cv2.COLOR_BGR2LAB)
cv2.imshow("lab",lab)
#-----Splitting the LAB image to different channels-------------------------
l, a, b = cv2.split(lab)
cv2.imshow('l_channel', l)
cv2.imshow('a_channel', a)
cv2.imshow('b_channel', b)
#-----Applying CLAHE to L-channel-------------------------------------------
clahe = cv2.createCLAHE(clipLimit=3.0, tileGridSize=(8,8))
cl = clahe.apply(l)
cv2.imshow('CLAHE output', cl)
#-----Merge the CLAHE enhanced L-channel with the a and b channel-----------
limg = cv2.merge((cl,a,b))
cv2.imshow('limg', limg)
#-----Converting image from LAB Color model to RGB model--------------------
final = cv2.cvtColor(limg, cv2.COLOR_LAB2BGR)
cv2.imshow('final', final)
#_____END_____#
Based on the great C++ example written by Bull, I was able to write this method for Android.
I have substituted "Core.extractChannel" for "Core.split". This avoids a known memory leak issue.
public void applyCLAHE(Mat srcArry, Mat dstArry) {
//Function that applies the CLAHE algorithm to "dstArry".
if (srcArry.channels() >= 3) {
// READ RGB color image and convert it to Lab
Mat channel = new Mat();
Imgproc.cvtColor(srcArry, dstArry, Imgproc.COLOR_BGR2Lab);
// Extract the L channel
Core.extractChannel(dstArry, channel, 0);
// apply the CLAHE algorithm to the L channel
CLAHE clahe = Imgproc.createCLAHE();
clahe.setClipLimit(4);
clahe.apply(channel, channel);
// Merge the the color planes back into an Lab image
Core.insertChannel(channel, dstArry, 0);
// convert back to RGB
Imgproc.cvtColor(dstArry, dstArry, Imgproc.COLOR_Lab2BGR);
// Temporary Mat not reused, so release from memory.
channel.release();
}
}
And call it like so:
public Mat onCameraFrame(CvCameraViewFrame inputFrame){
Mat col = inputFrame.rgba();
applyCLAHE(col, col);//Apply the CLAHE algorithm to input color image.
return col;
}
I have applied CLAHE to this channel and It looks good.
I calculate the perceived brightness channel of the image
a - > I change the image into HSV Colour space and I replace the V channel from the image by adding the CLAHE applied perceived brightness channel.
b -> I change the image into LAB Colour space. I replace the L channel from the image by adding the CLAHE applied perceived brightness channel.
Then I again convert the image into BGR format.
The python code for my steps
import cv2
import numpy as np
original = cv2.imread("/content/rqq0M.jpg")
def get_perceive_brightness(img):
float_img = np.float64(img) # unit8 will make overflow
b, g, r = cv2.split(float_img)
float_brightness = np.sqrt(
(0.241 * (r ** 2)) + (0.691 * (g ** 2)) + (0.068 * (b ** 2)))
brightness_channel = np.uint8(np.absolute(float_brightness))
return brightness_channel
perceived_brightness_channel = get_perceive_brightness(original)
clahe = cv2.createCLAHE(clipLimit=3.0, tileGridSize=(8,8))
clahe_applied_perceived_channel = clahe.apply(perceived_brightness_channel)
def hsv_equalizer(img, new_channel):
hsv = cv2.cvtColor(original, cv2.COLOR_BGR2HSV)
h,s,v = cv2.split(hsv)
merged_hsv = cv2.merge((h, s, new_channel))
bgr_img = cv2.cvtColor(merged_hsv, cv2.COLOR_HSV2BGR)
return bgr_img
def lab_equalizer(img, new_channel):
lab = cv2.cvtColor(original, cv2.COLOR_BGR2LAB)
l,a,b = cv2.split(lab)
merged_lab = cv2.merge((new_channel,a,b))
bgr_img = cv2.cvtColor(merged_hsv, cv2.COLOR_LAB2BGR)
return bgr_img
hsv_equalized_img = hsv_equalizer(original,clahe_applied_perceived_channel)
lab_equalized_img = lab_equalizer(original,clahe_applied_perceived_channel)
Output of the hsv_equalized_img
The output of the lab_equlized_img
You can try the following code:
#include "opencv2/opencv.hpp"
#include <iostream>
using namespace std;
using namespace cv;
int main(int argc, char** argv)
{
cout<<"Usage: ./executable input_image output_image \n";
if(argc!=3)
{
return 0;
}
int filterFactor = 1;
Mat my_img = imread(argv[1]);
Mat orig_img = my_img.clone();
imshow("original",my_img);
Mat simg;
cvtColor(my_img, simg, CV_BGR2GRAY);
long int N = simg.rows*simg.cols;
int histo_b[256];
int histo_g[256];
int histo_r[256];
for(int i=0; i<256; i++){
histo_b[i] = 0;
histo_g[i] = 0;
histo_r[i] = 0;
}
Vec3b intensity;
for(int i=0; i<simg.rows; i++){
for(int j=0; j<simg.cols; j++){
intensity = my_img.at<Vec3b>(i,j);
histo_b[intensity.val[0]] = histo_b[intensity.val[0]] + 1;
histo_g[intensity.val[1]] = histo_g[intensity.val[1]] + 1;
histo_r[intensity.val[2]] = histo_r[intensity.val[2]] + 1;
}
}
for(int i = 1; i<256; i++){
histo_b[i] = histo_b[i] + filterFactor * histo_b[i-1];
histo_g[i] = histo_g[i] + filterFactor * histo_g[i-1];
histo_r[i] = histo_r[i] + filterFactor * histo_r[i-1];
}
int vmin_b=0;
int vmin_g=0;
int vmin_r=0;
int s1 = 3;
int s2 = 3;
while(histo_b[vmin_b+1] <= N*s1/100){
vmin_b = vmin_b +1;
}
while(histo_g[vmin_g+1] <= N*s1/100){
vmin_g = vmin_g +1;
}
while(histo_r[vmin_r+1] <= N*s1/100){
vmin_r = vmin_r +1;
}
int vmax_b = 255-1;
int vmax_g = 255-1;
int vmax_r = 255-1;
while(histo_b[vmax_b-1]>(N-((N/100)*s2)))
{
vmax_b = vmax_b-1;
}
if(vmax_b < 255-1){
vmax_b = vmax_b+1;
}
while(histo_g[vmax_g-1]>(N-((N/100)*s2)))
{
vmax_g = vmax_g-1;
}
if(vmax_g < 255-1){
vmax_g = vmax_g+1;
}
while(histo_r[vmax_r-1]>(N-((N/100)*s2)))
{
vmax_r = vmax_r-1;
}
if(vmax_r < 255-1){
vmax_r = vmax_r+1;
}
for(int i=0; i<simg.rows; i++)
{
for(int j=0; j<simg.cols; j++)
{
intensity = my_img.at<Vec3b>(i,j);
if(intensity.val[0]<vmin_b){
intensity.val[0] = vmin_b;
}
if(intensity.val[0]>vmax_b){
intensity.val[0]=vmax_b;
}
if(intensity.val[1]<vmin_g){
intensity.val[1] = vmin_g;
}
if(intensity.val[1]>vmax_g){
intensity.val[1]=vmax_g;
}
if(intensity.val[2]<vmin_r){
intensity.val[2] = vmin_r;
}
if(intensity.val[2]>vmax_r){
intensity.val[2]=vmax_r;
}
my_img.at<Vec3b>(i,j) = intensity;
}
}
for(int i=0; i<simg.rows; i++){
for(int j=0; j<simg.cols; j++){
intensity = my_img.at<Vec3b>(i,j);
intensity.val[0] = (intensity.val[0] - vmin_b)*255/(vmax_b-vmin_b);
intensity.val[1] = (intensity.val[1] - vmin_g)*255/(vmax_g-vmin_g);
intensity.val[2] = (intensity.val[2] - vmin_r)*255/(vmax_r-vmin_r);
my_img.at<Vec3b>(i,j) = intensity;
}
}
// sharpen image using "unsharp mask" algorithm
Mat blurred; double sigma = 1, threshold = 5, amount = 1;
GaussianBlur(my_img, blurred, Size(), sigma, sigma);
Mat lowContrastMask = abs(my_img - blurred) < threshold;
Mat sharpened = my_img*(1+amount) + blurred*(-amount);
my_img.copyTo(sharpened, lowContrastMask);
imshow("New Image",sharpened);
waitKey(0);
Mat comp_img;
hconcat(orig_img, sharpened, comp_img);
imwrite(argv[2], comp_img);
}
Check here for more details.
