Feature detection
In the first two parts of this chapter, we focused on preparing and enhancing images, with all steps aimed at making the pixel data cleaner and more consistent. However, we will go a step further in this section by identifying meaningful structures and converting them into numerical values we can analyze. Once we can determine the numbers, we can compute simple measurements that use these numeric summaries, often called features, in further follow-up analysis.
We will start with the edges.
Edges
When we view an image, our eyes naturally focus on boundaries, which are the outlines of objects and the background. In image processing, these boundaries are known as edges. Essentially, an edge is a point where pixel values change rapidly over a small area. Detecting these changes creates a map of object boundaries, which serves as a helpful starting point for tasks such as shape detection, object segmentation, or data preparation for measurement and OCR.
...
