Introduction
In the previous tutorial, we saw how to change between different color spaces. Now, in this tutorial we will see yet another functionality of OpenCV, related to colors – colormaps. The colormaps provide us with the ability to map the value of one color to another color’s value and change it accordingly. For example, we can map the value of color (255, 0, 0) which is blue to value (0, 0, 255) which is red and this can change all the blue pixels which have the value (255, 0, 0) to red ones. This could be particularly useful if we want to change the colors of an image.
However, for colorful images this could require a lot of effort to make the mapping. This is because we need to have a mapping for each combination of the three colors – red, green and blue. And this is a huge amount of combinations. It is far easier to work with grayscale images, which only have values from zero to 255. With this functionality, we can create the so-called pseudo colorization. We can make the grayscale image look colorful. However, we will not get quite the desired result. If we really want to colorize an old grayscale image, we will have to use deep learning and other machine learning approaches. We will see this in later tutorials.
Colorization using pre-defined colormaps
In the first example, we will see how to apply colorization using pre-defined colormaps. OpenCV comes with several such maps. We will take one grayscale image and apply on it some of them to see the results. Below is the code.
// Colorize grayscale image using predefined colormaps
void colorizeGrayscaleImagePredefinedMap()
{
// Path to the input image
std::string l_pathToInputImage{ "../Resources/youngster.jpg" };
// Create an object to hold the image data of the first image
Mat l_image;
// Read the image date from a file with no change to color scheme
l_image = imread(l_pathToInputImage, IMREAD_COLOR);
// Check if we have read the first image data correctly
if (!l_image.data)
{
std::cout << "No image data \n";
return;
}
// Create the output image matrix
Mat l_outputImage;
// Colorize the image using the predefined color map
applyColorMap(l_image, l_outputImage, COLORMAP_CIVIDIS);
// Display the input image
namedWindow("Input", WINDOW_NORMAL);
cv::imshow("Input", l_image);
// Display the result image
namedWindow("Result", WINDOW_NORMAL);
cv::imshow("Result", l_outputImage);
}Again, we load one image from a file. Notice that this time we use IMREAD_COLOR as a read mode in order to read the pixels in all three channels, to which we will apply the color mapping.
In order to apply a colormap, we call the function applyColorMap(). Its first argument is the input image matrix, the second arguments is the output matrix. The third argument is one enum, which defines what pre-defined colormaps to use. Check here the available values. It also shows each colormap’s colors.
Finally, we display the result. Here it is how it looks on my side. You can go ahead and experiment with each available colormap to see what results you will get.
You can also try to apply colormap to a colorful image. I did this. You can see the result below.
Colorization using user-defined colormaps
The provided pre-defined colormaps are a finite number. OpenCV also provide you with a way to define your own colormaps. We will use the overloaded version of the applyColorMap() function. Here is an example code on how to do it.
// Colorize grayscale image using user-defined colormaps
void colorizeGrayscaleImageUserDefinedMap()
{
// Path to the input image
std::string l_pathToInputImage{ "../Resources/youngBlood.jpg" };
// Create an object to hold the image data of the first image
Mat l_image;
// Read the image date from a file as a three channel image
l_image = imread(l_pathToInputImage, IMREAD_COLOR);
// Check if we have read the first image data correctly
if (!l_image.data)
{
std::cout << "No image data \n";
return;
}
// Create the output image matrix
Mat l_outputImage;
// Create the colormap
uchar l_data[] = {
0, 0, 0, // 0
3, 0, 0, // 1
6, 0, 0, // 2
9, 0, 0, // 3
12, 0, 0, // 4
15, 0, 0, // 5
18, 0, 0, // 6
21, 0, 0, // 7
24, 0, 0, // 8
27, 0, 0, // 9
30, 0, 0, // 10
33, 0, 0, // 11
36, 0, 0, // 12
39, 0, 0, // 13
42, 0, 0, // 14
45, 0, 0, // 15
48, 0, 0, // 16
51, 0, 0, // 17
54, 0, 0, // 18
57, 0, 0, // 19
60, 0, 0, // 20
0, 3, 0, // 21
0, 6, 0, // 22
0, 9, 0, // 23
0, 12, 0, // 24
0, 15, 0, // 25
0, 18, 0, // 26
0, 21, 0, // 27
0, 24, 0, // 28
0, 27, 0, // 29
0, 30, 0, // 30
3, 3, 0, // 31
3, 6, 0, // 32
3, 9, 0, // 33
3, 12, 0, // 34
3, 15, 0, // 35
3, 18, 0, // 36
3, 21, 0, // 37
3, 24, 0, // 38
3, 27, 0, // 39
3, 30, 0, // 40
6, 3, 0, // 41
6, 6, 0, // 42
6, 9, 0, // 43
6, 12, 0, // 44
6, 15, 0, // 45
6, 18, 0, // 46
6, 21, 0, // 47
6, 24, 0, // 48
6, 27, 0, // 49
6, 30, 0, // 50
9, 3, 0, // 51
9, 6, 0, // 52
9, 9, 0, // 53
9, 12, 0, // 54
9, 15, 0, // 55
9, 18, 0, // 56
9, 21, 0, // 57
9, 24, 0, // 58
9, 27, 0, // 59
9, 30, 0, // 60
18, 3, 0, // 61
18, 6, 0, // 62
18, 9, 0, // 63
18, 12, 0, // 64
18, 15, 0, // 65
18, 18, 0, // 66
18, 21, 0, // 67
18, 24, 0, // 68
18, 27, 0, // 69
18, 30, 0, // 70
30, 33, 0, // 71
30, 36, 0, // 72
30, 39, 0, // 73
30, 42, 0, // 74
30, 45, 0, // 75
30, 48, 0, // 76
30, 51, 0, // 77
30, 54, 0, // 78
30, 57, 0, // 79
30, 60, 0, // 80
30, 3, 18, // 81
30, 6, 21, // 82
30, 9, 24, // 83
30, 12, 27, // 84
30, 15, 30, // 85
30, 18, 33, // 86
30, 21, 36, // 87
30, 24, 39, // 88
30, 27, 42, // 89
30, 30, 45, // 90
30, 33, 48, // 91
30, 36, 51, // 92
30, 39, 54, // 93
30, 42, 57, // 94
30, 45, 60, // 95
30, 48, 63, // 96
30, 51, 66, // 97
30, 54, 69, // 98
30, 57, 72, // 99
30, 60, 75, // 100
54, 3, 18, // 101
54, 6, 21, // 102
54, 9, 24, // 103
54, 12, 27, // 104
54, 15, 30, // 105
54, 18, 33, // 106
54, 21, 36, // 107
54, 24, 39, // 108
54, 27, 42, // 199
60, 30, 45, // 110
60, 33, 48, // 111
60, 36, 51, // 112
60, 39, 54, // 113
60, 42, 57, // 114
60, 45, 60, // 115
60, 48, 63, // 116
60, 51, 66, // 117
60, 54, 69, // 118
60, 57, 72, // 119
60, 60, 75, // 120
54, 30, 18, // 121
54, 36, 21, // 122
54, 39, 24, // 123
54, 42, 27, // 124
54, 45, 30, // 125
54, 48, 33, // 126
54, 51, 36, // 127
54, 54, 39, // 128
54, 57, 42, // 129
60, 60, 45, // 130
90, 33, 48, // 131
90, 36, 51, // 132
90, 39, 54, // 133
90, 42, 57, // 134
90, 45, 60, // 135
90, 48, 63, // 136
90, 51, 66, // 137
90, 54, 69, // 138
90, 57, 72, // 139
90, 60, 75, // 140
120, 30, 60, // 141
120, 36, 66, // 142
120, 39, 72, // 143
120, 42, 78, // 144
120, 45, 84, // 145
120, 48, 90, // 146
120, 51, 96, // 147
120, 54, 108, // 148
120, 57, 120, // 149
128, 60, 45, // 150
128, 66, 48, // 151
128, 76, 51, // 152
128, 79, 54, // 153
128, 82, 57, // 154
128, 85, 60, // 155
128, 88, 63, // 156
128, 101, 66, // 157
128, 104, 69, // 158
128, 127, 72, // 159
128, 130, 75, // 160
130, 130, 60, // 161
130, 136, 66, // 162
130, 139, 72, // 163
130, 142, 78, // 164
130, 145, 84, // 165
130, 148, 90, // 166
130, 151, 96, // 167
130, 154, 108,// 168
130, 157, 120,// 169
148, 160, 45, // 170
148, 160, 48, // 171
148, 126, 51, // 172
148, 129, 54, // 173
148, 132, 57, // 174
148, 135, 60, // 175
148, 138, 63, // 176
148, 141, 66, // 177
148, 144, 69, // 178
148, 147, 72, // 179
148, 150, 75, // 180
150, 160, 60, // 181
150, 166, 66, // 182
150, 169, 72, // 183
150, 172, 78, // 184
150, 175, 84, // 185
150, 178, 90, // 186
150, 181, 96, // 187
150, 184, 108,// 188
150, 187, 120,// 189
168, 160, 75, // 190
168, 160, 78, // 191
168, 126, 81, // 192
168, 129, 84, // 193
168, 132, 87, // 194
168, 135, 90, // 195
168, 138, 93, // 196
168, 141, 96, // 197
168, 144, 99, // 198
168, 147, 122,// 199
168, 150, 125,// 200
180, 160, 60, // 201
180, 166, 66, // 202
180, 169, 72, // 203
180, 172, 78, // 204
180, 175, 84, // 205
180, 178, 90, // 206
180, 181, 96, // 207
180, 184, 108,// 208
180, 187, 120,// 209
196, 160, 75, // 210
196, 160, 78, // 211
196, 126, 81, // 212
196, 129, 84, // 213
196, 132, 87, // 214
196, 135, 90, // 215
196, 138, 93, // 216
196, 141, 96, // 217
196, 144, 99, // 218
196, 147, 122,// 219
196, 150, 125,// 220
200, 180, 60, // 221
200, 186, 66, // 222
200, 189, 72, // 223
200, 192, 78, // 224
200, 195, 84, // 225
200, 198, 90, // 226
200, 201, 96, // 227
200, 204, 108,// 228
200, 207, 120,// 229
216, 230, 120,// 230
216, 240, 120,// 231
216, 246, 120,// 232
216, 249, 122,// 233
216, 252, 122,// 234
216, 252, 132,// 235
216, 252, 132,// 236
216, 255, 145,// 237
216, 255, 148,// 238
216, 255, 156,// 239
216, 255, 160,// 240
230, 180, 60, // 241
230, 186, 66, // 242
230, 189, 72, // 243
230, 192, 78, // 244
230, 195, 84, // 245
230, 198, 90, // 246
230, 201, 96, // 247
230, 204, 108,// 248
230, 207, 120,// 249
236, 230, 150,// 250
236, 240, 150,// 251
236, 246, 150,// 252
242, 249, 162,// 253
242, 252, 162,// 254
242, 252, 172,// 255
};
Mat l_colorMap{ 256, 1, CV_8UC3, l_data };
// Colorize the image using the user-defined color map
applyColorMap(l_image, l_outputImage, l_colorMap);
// Display the input image
namedWindow("Input", WINDOW_NORMAL);
cv::imshow("Input", l_image);
// Display the result image
namedWindow("Result", WINDOW_NORMAL);
cv::imshow("Result", l_outputImage);
}As with the previous example, we again load the image in IMREAD_COLOR mode. Then it is time to define the color mapping. We need to define three values for each grayscale level from zero to 255. As you can see, it is quite a long list. In my case, I have defined random numbers. You can define some numbers, which will make sense to you. From the defined data, we create one colormap matrix.
Once we are done with defining the colormap it is time to call the applyColorMap() function. As I said, we call its overloaded version, which will take the colormap matrix as a third argument. It will apply the colormapping to the grayscale image we provide. We will display the result. You can see what the result on my side was. It is quite random.
You can also apply this random colormap to a color image. See what I have got on my side.
Conclusion
In this tutorial, we saw how to use colormaps to apply colorization on images. The colorization, which we achieved, is not a real one. It is pseudo colorization. If you want to achieve a real colorization, you need to use machine learning. Nevertheless, you can achieve quite impressive results. Try on your side all the available pre-defined colormaps and create your own.
The code is again available on GitHub and as an archive below.
Next, we will see how to use color changing and thresholding.
OpenCV Training
Passionate developer, loving husband and caring father. As an introvert programming is my life. I work as a senior software engineer in a big international company. My hobbies include playing computer games (mostly World of Warcraft), watching TV series and hiking.
