OpenCV or such visual libraries are used for dynamic object recognition, when you need to detect that something is a tree, or a rock, or water. And all those objects change dynamically, meaning they look slightly different all the time. So it's more complicated then just some color detection. But if your task is related to simple colors or static pictures, then you can easily do it with pixels only. You can also use text recognition, which is faster than object recognition.