Image object detection question

I am about to start on a university project that is not part of any courses. I was tasked with identifying objects on a arbitrary image (say number of circles in a image). I will be working mostly with 2D images. I don't have much experience with such image processing and was wondering if someone could give a guidance on where to start.

My initial google search gave me some keywords to build upon, such as OpenCV. I have some courses regarding signals and systems (fourier transform in parti开发者_开发知识库cular), lots of computer science related courses (data structures, algorithms, etc), math courses (calculus, linear algebra, etc), and others.

Can someone please give me a quick guide on what kind of things to look for and learn? Any kind of guidance would be greatly appreciated. Thanks!

---

Machine vision image processing is an extremely complex area. I recommend the book Image Processing, Analysis, and Machine Vision by Milan Sonka. An expensive tome, but unique in being oriented toward practical algorithms.

---

Read chapter 14 of Richard Szeleski's new book (PDF draft available) for an overview of current techniques.

There's also material from a short conference course here.

---

Some simple object such as Circles can be detected using Hough transform. If you need some more advanced object detection you can use Viola-Jones algorithm (see e.g. OpenCV or Rapidminer http://spl.utko.feec.vutbr.cz/en/image-processing-extension-for-rapidminer-5) or Histogram of Oriented Gradients (is planet to http://spl.utko.feec.vutbr.cz/en/image-processing-extension-for-rapidminer-5)