Generally speaking,the golden age of IT is the XXI. century, we are aiming to automation in all areas of life. The main aim of my thesis is to present the problem of pattern and shape recognition in the topic of machine vision.
After represeting the theoretical background. I’m going to describe the topic which was chosen as the theme of my thesis by designing and implementing high-level shape recognition modules, which function together as a program library. The solution is capable of detecting multiple shapes on all input images, such as triangle, parallelogram, rectangle, square, octogonal and circle. For the implementation I used PyCharm IDE and the modules were written in Python programing language. The standalone tests of the initial modules were provided by a dataset, which contains 11205 images and downloaded from Kaggle website.
I adapted the whole library into two directions, which was the second step of the implementation. I was able to test shapes of the traffic signs with the help of the dataset, called UAH, and detect the series of equal shapes, in which case the distance between the shapes is almost equal. Finally, I present the test results on the images of the dataset and formulate further development possibilities.