Color Filtering/Segmentation/Detection – HSV - Computer Vision

One of the first hands-on application that fledgling Computer Vision enthusiasts start with is color filtering/detection.

Why using HSV color space for this?

** In real world images there is always variations in the image color values due to various lightening conditions, shadows and, even due to noise added by the camera while clicking and subsequently processing the image.

• To over the above color variations, we will perform color detection in the HSV color space.

The whole idea is, we will be creating a binary mask where the white region will represent our target color (red in this case) and black will represent rest of the colors.

So we will be using the inRange() function to filter out the required color.

mask = cv2.inRange(src, lowerb, upperb)

• src: This will be src array, in our case it will be the target Hsv image
• lowerb: This is the lower boundary array, in our case it will be an array of 3 values , these values will be the lower range of hsv values.
• lowerb: This is the upper boundary array, in our case it will be an array of 3 values , these values will be the upper range of hsv values.

Now using inRange() function, we will filter all the colors that are in between the range of lower and upper boundaries. For example to capture red color we can filter out Hue in range 160-180 and the values and saturation can be in range 50-255. This will give us different variations of red both dark-bright and dull-pure red color variations are captured.

Ref link – http://hanzratech.in/2015/02/07/caveat-thresholding-hue-component.html

https://stackoverflow.com/questions/30331944/finding-red-color-in-image-using-python-opencv

How to get HSV values?

OpenCV halves the H values to fit the range [0,255], so H value instead of being in range [0, 360], is in range [0, 180]. S and V are still in range [0, 255].

Look at this site which gives you HSV values for any RGB value.

import cv2 import numpy as np greenBGR = np.uint8([[[0,255,0 ]]]) hsv_green = cv2.cvtColor(greenBGR,cv2.COLOR_BGR2HSV) print (hsv_green)

Output :

[[[ 60 255 255]]]

Now you take [H-10, 100,100] and [H+10, 255, 255] as lower bound and upper bound respectively.

In this blog post we will try to detect the rose flower in the image below using color detection:

In OpenCV, Hue has values from 0 to 180, Saturation and Value from 0 to 255. Thus, OpenCV uses HSV ranges between (0-180, 0-255, 0-255). In OpenCV, the H values 179, 178, 177 and so on are as close to the true RED as H value 1, 2, 3 and so on.

HSV color map:

From the above HSV color map, to find a color, usually just look up for the range of H and S, and set v in range (20, 255).

To find the orange color, we look up for the map, and find the best range: H :[10, 25], S: [100, 255], and V: [20, 255] So the mask is cv2.inRange(hsv,(10, 100, 20), (25, 255, 255) )

Our rose flower is predominantly red, so we will set the cv2.inrange function with the range of HSV values of red color.The HSV values for true RED are (0, 255, 255) and to accommodate for color variations, we will consider a range of HSV values for the red color.

• The red color, in OpenCV, has the hue values approximately in the range of 0 to 10 and 160 to 180.

We will use the cv2.inRange to generate the mask that has a value of 255 for pixels where the HSV values fall within the specified color range and a value of 0 for pixels whose values don’t lie in this interval. The mask values are either 255 or 0. 255 represents color pixels and 0 represents non color pixels.

Full code :

import cv2 import numpy as np image = cv2.imread('rose.jpg') cv2.imshow("Original", image) result = image.copy() image = cv2.cvtColor(image, cv2.COLOR_BGR2HSV) # lower boundary RED color range values; Hue (0 - 10) lower1 = np.array([0, 100, 20]) upper1 = np.array([10, 255, 255]) # upper boundary RED color range values; Hue (160 - 180) lower2 = np.array([160,100,20]) upper2 = np.array([179,255,255]) lower_mask = cv2.inRange(image, lower1, upper1) upper_mask = cv2.inRange(image, lower2, upper2) full_mask = lower_mask + upper_mask; result = cv2.bitwise_and(result, result, mask=full_mask) cv2.imshow('mask', full_mask) cv2.imshow('result', result) cv2.waitKey(0) cv2.destroyAllWindows()

Output :

Another example :

import cv2 import matplotlib.pyplot as plt import numpy as np image = cv2.imread('media/M3/stop.jpg') # Converting the image to hsv hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV) # define range of red color in HSV lower_red = np.array([160,50,50]) upper_red = np.array([180,255,255]) # Threshold the HSV image using inRange function to get only red colors mask = cv2.inRange(hsv, lower_red, upper_red) plt.figure(figsize=[13,13]) plt.subplot(121);plt.imshow(image[:,:,::-1]);plt.title("Original Image",fontdict={'fontsize': 25});plt.axis('off'); plt.subplot(122);plt.imshow(mask, cmap='gray');plt.title("Mask of red Color",fontdict={'fontsize': 25});plt.axis('off');

Output :

Now we can use this Mask and pass into the bitwise_and() function along with the original image and it will give the part of the image which has red color.

res = cv2.bitwise_and(image,image, mask= mask) plt.figure(figsize=[13,13]) plt.imshow(res[:,:,::-1]);plt.title("Red part of the Image",fontdict={'fontsize':35});plt.axis('off');

Output :

Capture a Colored object in Real time / Color Detection / Find Color object

import cv2 import matplotlib.pyplot as plt import numpy as np cap = cv2.VideoCapture(0) while(1): # Take each frame _, frame = cap.read() # Convert BGR to HSV hsv = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV) # define range of red color in HSV lower_red = np.array([160,50,50]) upper_red = np.array([180,255,255]) #Threshold the HSV image to get only red colors mask = cv2.inRange(hsv, lower_red, upper_red) # Bitwise-AND mask and original image res = cv2.bitwise_and(frame,frame, mask= mask) # we are just adding 2 more channels on the mask so we can stack it along other images mask_3 = cv2.cvtColor(mask, cv2.COLOR_GRAY2BGR) # stacking up all three images together stacked = np.hstack((mask_3,frame,res)) cv2.imshow('Result',cv2.resize(stacked,None,fx=0.8,fy=0.8)) k = cv2.waitKey(1) & 0xFF if k == 27: break cv2.destroyAllWindows() cap.release()

Share this:

• Twitter
• Facebook

Like Loading...

Related