A few suggestions to improve the writeup:
1. Consider going into more depth on the workings of the image processing techniques. The ones you have chosen are some of the more readily-accessible and, to the uninitiated, can be a delight to learn (esp the Hough Transform, imo). This will also help cement things in your mind and be useful when it comes to exam time.
2. Feature selection is obviously the most critical part of the process - consider going into more detail. (Un)Fortunately you got it right first time, so you couldn't document a process of trial, error, insight and finally success. In lieu of this you should at least be able to elucidate on the thought processes that led you to choose the features you did. If you are able to make a prediction that is subsequently realised then that demonstrates you are intelligently selecting features and understanding the consequences, rather than just getting lucky.
3. Your description of k-Nearest Neighbours is a little faulty - possibly the result of you trying to phrase what is essentially a simple concept in your own words? The user doesn't pre-define some set 'search radius' - the algorithm simply finds the k nearest neighbours of a given point. If your point is some outlier then the 'search radius' will be much larger then had you selected a more 'typical' point in the middle of a densely-populated area - for the same value of k.