Detect object in an image using OpenCV Python.
As example we will use a picture of a green apple:

Initial picture on the left. Processed picture on the right. Green ellipse indicates borders of an apple.

Firstly, we need to convert image from BGR color scheme to RGB. We do it using cv2.cvtColor() function
Resize image to 700x700 to have standardized representation of image
To remove noise and redundant details from image we will use cv2.GaussianBlur() function with kernel size 5x5 and zero standard deviation
After that we will convert image to HSV color scheme. At this point image will look like this:

We need to distinguish apple from the rest of the image, so that we will use masks.
We will filter image by color and then by brightness using cv2.inRange() function, after that masks will be combined:

Create ellipse shaped kernel of size 15x15 using cv2.getStructuringElement() function.
Perform closing and then opening operations using cv2.morphologyEx() with kernel specified above to remove noise and to close small holes inside foreground object.

After having preprocessed mask find contours with cv2.findContours() and then detect the biggest one.
As we found contours, we can combine contour, mask and original image.
Finally, change color scheme back from RGB to BGR and display final image: