python-3.xopencvcolor-detectionpython

Detect if an object has a specific colour out of 3 colours using opencv (cv2 in python3)

I am currently working on a program that can detect if a red object is kept in my room or is is there a blue object. The rest of my surroundings is either white or black. I have tried to minimize the variation of light in my room.

I have successfully created a mask around the object given a certain hue range. I want my program to print for me :

1) "Red"- If there is a red object

2) "Blue"- If there is a blue object

I don't know how to proceed. Following is my program that cretes the mask around the object that is blue. I have given the hue range of a few other colours also. So that you can try it.

The program:

import cv2
import numpy as np

cam = cv2.VideoCapture(1)

while True:
    _, frame = cam.read()

    denoised = cv2.GaussianBlur(frame, (31, 31), 35)
    hsv = cv2.cvtColor(denoised, cv2.COLOR_BGR2HSV)


    lower_blue = np.array([110, 50, 50])
    upper_blue = np.array([160, 255, 255])

    mask = cv2.inRange(hsv, lower_red, upper_red)
    res = cv2.bitwise_and(frame, frame, mask=mask)

    cv2.imshow('frame', frame)
    #cv2.imshow('mask', mask)
    cv2.imshow('res', res)

    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cam.release()
cv2.destroyAllWindows()

Hue of different colours (I am not sure about the hue of red because it doesn't work for some colours- I have tried a few solutions from Stackoverflow):

lower_red = np.array([0, 100, 100])
upper_red = np.array([0, 255, 255])

lower_yellow = np.array([15, 210, 20])
upper_yellow = np.array([35, 255, 255])

lower_green = np.array([29, 86, 6])
upper_green = np.array([64, 255, 2555])

lower_orange = np.array([10, 100, 20])
upper_orange = np.array([20,255,255])

Following are some sample images that you can experiment with:

Blue Image

Red Image

enter image description here

Solution

  • Your approach is correct in a way. But to determine the color of a specific region of the image, you need to calculate the Euclidean distance between the known dataset of the colors and the L*a*b averages of the region.

    1. Detect a specific region of interest, for which you need the color.

    2. Refer the following code to determine the color within the region of interest.

      class ColorLabeler:
    def __init__(self):
        # initialize the colors dictionary, containing the color
        # name as the key and the RGB tuple as the value
        colors = OrderedDict({
            "red": (255, 0, 0),
            "green": (0, 255, 0),
            "blue": (0, 0, 255)})

        # allocate memory for the L*a*b* image, then initialize
        # the color names list
        self.lab = np.zeros((len(colors), 1, 3), dtype="uint8")
        self.colorNames = []

        # loop over the colors dictionary
        for (i, (name, rgb)) in enumerate(colors.items()):
            # update the L*a*b* array and the color names list
            self.lab[i] = rgb
            self.colorNames.append(name)

        # convert the L*a*b* array from the RGB color space
        # to L*a*b*
        self.lab = cv2.cvtColor(self.lab, cv2.COLOR_RGB2LAB)

def label(self, image, c):
    # construct a mask for the contour, then compute the
    # average L*a*b* value for the masked region
    mask = np.zeros(image.shape[:2], dtype="uint8")
    cv2.drawContours(mask, [c], -1, 255, -1)
    mask = cv2.erode(mask, None, iterations=2)
    mean = cv2.mean(image, mask=mask)[:3]

    # initialize the minimum distance found thus far
    minDist = (np.inf, None)

    # loop over the known L*a*b* color values
    for (i, row) in enumerate(self.lab):
        # compute the distance between the current L*a*b*
        # color value and the mean of the image
        d = dist.euclidean(row[0], mean)

        # if the distance is smaller than the current distance,
        # then update the bookkeeping variable
        if d < minDist[0]:
            minDist = (d, i)

    # return the name of the color with the smallest distance
    return self.colorNames[minDist[1]]