Machine vision application in laboratory environment

Zadatak završnog rada obuhvaća primjenu strojnog vida u laboratorijskom okruženju. U uvodnom dijelu opisane su karakteristike robota, industrijskih kamera i strojnog vida u svrhu lakšeg praćenja rada. Također metode obrade i modeli prikaza boje te način njihovog konvertiranja iz jednog modela u drugi. Pod obradama slika su opisani pojedini filtri slike i procedure uklanjanja smetnji (šumova). Također su objašnjene pojedine transformacije slike koje se koriste kako bi se uočili krugovi i pronašli rubovi na slikama. Praktični dio rada obuhvaća pronalaženje definiranog izratka s otvorima na transportnoj traci, kontroliranje broja otvora na izratku te sortiranje istih pomoću robotske ruke. Izradak na transportnoj traci se preko kamere uočava HSV modelom prikaza boje. Pomoću gotovih funkcija u OpenCV-u (cv2.Canny() i cv2.findContours()) pronalaze se otvori (zatvorene konture) na izratku. Zatvorene konture se prebrojavaju i izradci se pomoću robotske ruke sortiraju na ispravne i neispravne izratke. Pojedini programski kodovi za izvršavanje praktičnog dijela sadržane su na kraju završnog rada.The final paper assigment involves the use of machine vision in the laboratory environment. The introductory section describes the characteristics of robots, industrial cameras and machine vision for the purpose of easier monitoring of work. Also processing methods and color display models and the way of converting them from one model to another. Under color processing are described individual image filters and disturbance removal procedures (noise). Some image transformations that are used to spot circles and find the edges in the image are also explained.Practical part of the work involves finding a defined workpiece with openings on the conveyor, controlling the number of openings on the workpiece and sorting them using a robotic hand. The workpiece on the conveyor is detected by the HSV model of the color space over the camera. With open functions in OpenCV (cv2.Canny() and cv2.findContours()), openings (closed contours) are found on the work. Closed contours are counted, and the workpieces are sorted by robotic hands on the correct and defective workpieces. Some program codes for performing a practical part are included at the end of the final work