You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
- In this tutorial, you will learn Simple thresholding, Adaptive thresholding, Otsu's thresholding
8
-
etc.
9
-
- You will learn these functions : **cv.threshold**, **cv.adaptiveThreshold** etc.
7
+
- In this tutorial, you will learn Simple thresholding, Adaptive thresholding and Otsu's thresholding.
8
+
- You will learn the functions **cv.threshold** and **cv.adaptiveThreshold**.
10
9
11
10
Simple Thresholding
12
11
-------------------
13
12
14
-
Here, the matter is straight forward. If pixel value is greater than a threshold value, it is
15
-
assigned one value (may be white), else it is assigned another value (may be black). The function
16
-
used is **cv.threshold**. First argument is the source image, which **should be a grayscale
17
-
image**. Second argument is the threshold value which is used to classify the pixel values. Third
18
-
argument is the maxVal which represents the value to be given if pixel value is more than (sometimes
19
-
less than) the threshold value. OpenCV provides different styles of thresholding and it is decided
20
-
by the fourth parameter of the function. Different types are:
13
+
Here, the matter is straight forward. For every pixel, the same threshold value is applied.
14
+
If the pixel value is smaller than the threshold, it is set to 0, otherwise it is set to a maximum value.
15
+
The function **cv.threshold** is used to apply the thresholding.
16
+
The first argument is the source image, which **should be a grayscale image**.
17
+
The second argument is the threshold value which is used to classify the pixel values.
18
+
The third argument is the maximum value which is assigned to pixel values exceeding the threshold.
19
+
OpenCV provides different types of thresholding which is given by the fourth parameter of the function.
20
+
Basic thresholding as described above is done by using the type cv.THRESH_BINARY.
21
+
All simple thresholding types are:
21
22
22
23
- cv.THRESH_BINARY
23
24
- cv.THRESH_BINARY_INV
24
25
- cv.THRESH_TRUNC
25
26
- cv.THRESH_TOZERO
26
27
- cv.THRESH_TOZERO_INV
27
28
28
-
Documentation clearly explain what each type is meant for. Please check out the documentation.
29
+
See the documentation of the types forthe differences.
29
30
30
-
Two outputs are obtained. First one is a **retval** which will be explained later. Second output is
31
-
our**thresholded image**.
31
+
The method returns two outputs.
32
+
The first is the threshold that was used and the second output is the**thresholded image**.
32
33
33
-
Code :
34
+
This code compares the different simple thresholding types:
34
35
@code{.py}
35
36
import cv2 as cv
36
37
import numpy as np
@@ -53,34 +54,31 @@ for i in xrange(6):
53
54
54
55
plt.show()
55
56
@endcode
56
-
@note To plot multiple images, we have used plt.subplot() function. Please checkout Matplotlib docs
57
-
for more details.
57
+
@note To plot multiple images, we have used the plt.subplot() function. Please checkout the matplotlib docs for more details.
58
58
59
-
Result is given below :
59
+
The code yields this result:
60
60
61
61
62
62
63
63
Adaptive Thresholding
64
64
---------------------
65
65
66
-
In the previous section, we used a global value as threshold value. But it may not be good in all
67
-
the conditions where image has different lighting conditions in different areas. In that case, we go
68
-
for adaptive thresholding. In this, the algorithm calculate the threshold for a small regions of the
69
-
image. So we get different thresholds for different regions of the same image and it gives us better
70
-
results for images with varying illumination.
66
+
In the previous section, we used one global value as a threshold.
67
+
But this might not be good in all cases, e.g. if an image has different lighting conditions in different areas.
68
+
In that case, adaptive thresholding thresholding can help.
69
+
Here, the algorithm determines the threshold for a pixel based on a small region around it.
70
+
So we get different thresholds for different regions of the same image which gives better results for images with varying illumination.
71
71
72
-
It has three ‘special’ input params and only one output argument.
72
+
Additionally to the parameters described above, the method cv.adaptiveThreshold three input parameters:
73
73
74
-
**Adaptive Method**- It decides how thresholding value is calculated.
75
-
- cv.ADAPTIVE_THRESH_MEAN_C : threshold value is the mean of neighbourhood area.
76
-
- cv.ADAPTIVE_THRESH_GAUSSIAN_C : threshold value is the weighted sum of neighbourhood
77
-
values where weights are a gaussian window.
74
+
The **adaptiveMethod** decides how the threshold value is calculated:
75
+
- cv.ADAPTIVE_THRESH_MEAN_C: The threshold value is the mean of the neighbourhood area minus the constant **C**.
76
+
- cv.ADAPTIVE_THRESH_GAUSSIAN_C: The threshold value is a gaussian-weighted sum of the neighbourhood
77
+
values minus the constant **C**.
78
78
79
-
**Block Size** - It decides the size of neighbourhood area.
79
+
The **blockSize** determines the size of the neighbourhood area and **C** is a constant that is subtracted from the mean or weighted sum of the neighbourhood pixels.
80
80
81
-
**C** - It is just a constant which is subtracted from the mean or weighted mean calculated.
82
-
83
-
Below piece of code compares global thresholding and adaptive thresholding for an image with varying
81
+
The code below compares global thresholding and adaptive thresholding for an image with varying
84
82
illumination:
85
83
@code{.py}
86
84
import cv2 as cv
@@ -106,33 +104,30 @@ for i in xrange(4):
106
104
plt.xticks([]),plt.yticks([])
107
105
plt.show()
108
106
@endcode
109
-
Result:
107
+
Result:
110
108
111
109
112
110
113
-
Otsu’s Binarization
111
+
Otsu's Binarization
114
112
-------------------
115
113
116
-
In the first section, I told you there is a second parameter **retVal**. Its use comes when we go
117
-
for Otsu’s Binarization. So what is it?
118
-
119
-
In global thresholding, we used an arbitrary value for threshold value, right? So, how can we know a
120
-
value we selected is good or not? Answer is, trial and error method. But consider a **bimodal
121
-
image** (*In simple words, bimodal image is an image whose histogram has two peaks*). For that
122
-
image, we can approximately take a value in the middle of those peaks as threshold value, right ?
123
-
That is what Otsu binarization does. So in simple words, it automatically calculates a threshold
124
-
value from image histogram for a bimodal image. (For images which are not bimodal, binarization
125
-
won’t be accurate.)
126
-
127
-
For this, our cv.threshold() function is used, but pass an extra flag, cv.THRESH_OTSU. **For
128
-
threshold value, simply pass zero**. Then the algorithm finds the optimal threshold value and
129
-
returns you as the second output, retVal. If Otsu thresholding is not used, retVal is same as the
130
-
threshold value you used.
131
-
132
-
Check out below example. Input image is a noisy image. In first case, I applied global thresholding
133
-
for a value of 127. In second case, I applied Otsu’s thresholding directly. In third case, I
134
-
filtered image with a 5x5 gaussian kernel to remove the noise, then applied Otsu thresholding. See
135
-
how noise filtering improves the result.
114
+
In global thresholding, we used an arbitrary chosen value as a threshold.
115
+
In contrast, Otsu's method avoids having to choose a value and determines it automatically.
116
+
117
+
Consider an image with only two distinct image values (*bimodal image*), where the histogram would only consist of two peaks.
118
+
A good threshold would be in the middle of those two values.
119
+
Similarly, Otsu's method determines an optimal global threshold value from the image histogram.
120
+
121
+
In order to do so, the cv.threshold() function is used, where cv.THRESH_OTSU is passed as an extra flag.
122
+
The threshold value can be chosen arbitrary.
123
+
The algorithm then finds the optimal threshold value which is returned as the first output.
124
+
125
+
Check out the example below.
126
+
The input image is a noisy image.
127
+
In the first case, global thresholding with a value of 127 is applied.
128
+
In the second case, Otsu's thresholding is applied directly.
129
+
In the third case, the image is first filtered with a 5x5 gaussian kernel to remove the noise, then Otsu thresholding is applied.
0 commit comments