The 20th century was characterized by global economic and social transformation, which is now referred to as the second industrial revolution. One of the procedure of this transformation was the automation; through the session, step by step, the role of people was taken over by automatic systems. This kind of transformation could not avoid the agriculture either. Today, the processing of one of our agricultural products, the pollen is mostly automated. The pollen is also sold as a dietary supplement, thus ensuring a good quality pollen is very important. However, the pollen cannot be completely cleaned of various impurities by filters, the maximum purity can be achieved manually using tweezers. It is a difficult, time consuming and not very effective method, therefore the demand for automation of pollen treatment is given.
Through the thesis an image processing system was designed to determine the location of the inappropriate soiling. Various methods have been tested, and then I selected the most effective method by evaluating the results. After locating the contamination, the parameters were given to a robotic handling device, which pick out the impurities, so the cleaning process for pollen becomes easier and more effective.