OpenCV: Floor removal from image-CodePudding

I am trying to detect an object on the floor (the floor has a pattern). The object is the cup-like thing in the middle.

I tried to use Canny and Sobel edge detection, however it detects some parts of the floor too. I also tried to HSV color filtering, however it is prone to the room's lighting. SIFT couldn't detect the object too.

How can I remove the floor from the image?

Thanks

CodePudding user response：

Here is how you can use the Canny edge detector:

import cv2
import numpy as np

def process(img):
    img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    img_blur = cv2.GaussianBlur(img_gray, (5, 5), 6)
    img_canny = cv2.Canny(img_blur, 119, 175)
    kernel = np.ones((9, 3))
    img_dilate = cv2.dilate(img_canny, kernel, iterations=7)
    return cv2.erode(img_dilate, kernel, iterations=7)

img = cv2.imread("lPNAJ.jpg")
contours = cv2.findContours(process(img), cv2.RETR_TREE, cv2.CHAIN_APPROX_NONE)[0]
cnt = sorted(contours, key=cv2.contourArea)[-2]
cv2.drawContours(img, [cv2.convexHull(cnt)], -1, (0, 0, 255), 2)
cv2.imshow("result", img)
cv2.waitKey(0)
cv2.destroyAllWindows()

Output for the first image as input:

Output for the second image as input:

Note that the code takes the second largest contour present in the image, as the largest one belongs to the table on the right.