Python Image Processing
Undergraduate projects, 2017
Python Image Processing project
A simple camera like webcams will give us digital images that we could save them or process with our codes (algorithm) in python, MATLAB, C#, Java and … . Common needs in image processing include Classification, Feature extraction, Multi Scale Signal analysis, Pattern Recognition, Projection and … or simple tasks like rotating, flipping, cropping and … .
OpenCV (Open Source Computer Vision Library) is an open source computer vision and machine learning software library. OpenCV was built to provide a common infrastructure for computer vision applications and to accelerate the use of machine perception in the commercial products. Being a BSD-licensed product, OpenCV makes it easy for businesses to utilize and modify the code.
First we need to get skills in python or other programming languages like python, I’d chosen python for massive online documentations and help forums and I hope this programing language will get more popular than before as powerful modules keep getting professional and complex.
Eye & Face detection/tracker
Here I’d written some simple codes to get what I had needed in my machines to do some command if some features appears in camera. One of the amazing codes is to recognizing or detecting some features like eyes or faces in camera; results will be like below pictures that were taken by my laptop’s webcam. (the codes will be uploaded as soon as possible.)
Medical/Herbal plant detection/tracker
My previous project was matching or finding the correct species of plants or flowers by processing some databases in pharmacology to find best match as a picture is taken by smart phones or pictures from small organisms like bacteria under microscopes. In this area I’d to study their features and conditions in any circumstances. some of them was the color, vessels, leaves and … to know which feature is a key feature to match or work on.
Mentha pulegium plant is plentiful in Azerbaijan-Iran and my village “Gündoğdu” and we usually use it. The results were like below in our primary codes.
