Read the articles on computer vision on docs.wpilib.org. Since we'll be using OpenCV in C++, pay attention to the C++ code examples on those pages.
For an example of using vision for motion control, watch this video by FRC team 254. It gives advice on building a robust vision pipeline and how to combine it with other feedback controllers on the robot while compensating for camera and processing latency.
You should now have an understanding of the basic principles involved in computer vision. However, it takes practice to gain an intuition for what pipeline operations to apply to find the desired features in an image. GRIP can help here by providing a drag-and-drop interface that shows the output of intermediate processing steps. It can be downloaded from here.