Algoritem za samodejno ostrenje slike za industrijske kamere : diplomsko delo visokošolskega strokovnega študija

Klenovšek, Lovro (author), Pirc, Matija (mentor)

V diplomski nalogi je predstavljen razvoj metode samodejnega ostrenja, ki se uporablja s pomočjo samostojnih objektivov, ki služijo kot dodatek za kamero. V podjetju se je pojavila zahteva za optimizacijo algoritma samodejnega ostrenja, saj je bil dotedanji algoritem nezanesljiv. Samodejno ostrenje bo tako služilo kot pripomoček končnim uporabnikom, da bodo lahko sliko izostrili s pritiskom na gumb. Izdelal sem gonilnik za izračun vrednosti ostrine slike, gonilnik za premik leče s pomočjo motorja in algoritem za samodejno ostrenje slik. Za izračun vrednosti ostrine slike sem uporabil Brennerjev gradient, kar je eden izmed pasivnih načinov samodejnega ostrenja, ki je primeren za nepremične slike. Za premik leče z motorjem sem uporabljal PID regulacijo in pulzno širinsko modulacijo. Sistem je sestavljen iz grafično procesne enote, objektiva, kamere in krmilnega vezja. Celoten projekt je bil napisan v programskem jeziku C s pomočjo programa Keil. Skozi izdelavo projekta sem dosegel vse zastavljene cilje in ga brez težav z uspehom dokončal. Program, ki sem ga spisal, se tako lahko uporablja s katerem koli objektivom, sliko izostri ne glede na to ali je le ta barvna ali črno bela, njegovega delovanja pa naj nebi zmotili tudi nobeni zunanji dejavniki kot so npr. vremenske razmere. Program je zdaj v porabi podjetja DATCON, kot del prodajne ponudbe.

• Language: Slovenian
• Keywords: Objektiv, Samodejno ostrenje, PID regulacija, PWM, Brennerjev gradient
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FE - Faculty of Electrical Engineering
• Place of publishing: Ljubljana
• Publisher: [L. Klenovšek]
• Year: 2021
• Number of pages: XIV, 46 str.
• PID: 20.500.12556/RUL-128523
• UDC: 004.4:681.5:681.7(043.2)
• COBISS.SI-ID: 71244547
• Publication date in RUL: 16.07.2021

Secondary language

• Language: English
• Title: Autofocus algorithm for industrial cameras

Abstract

The thesis presents the development of the autofocus method, which is used with independent lenses that serve as an accessory for the camera. There was a demand in the company to optimize the autofocus algorithm, as the previous algorithm was unreliable. Autofocus will serve as an aid to end users so that they can focus the image at the touch of a button. I wrote a driver to calculate the image sharpness value, a driver to move the lens using a motor and an algorithm to automatically focus the images. I used the Brenner gradient to calculate the image sharpness value, which is one of the passive autofocus modes suitable for still images. To move the lens with the motor I used a PID control algorithm and a pulse width modulation. The system consists of a graphic processing unit, a lens, a camera and a control circuit. The entire project was written in the C programming language using the Keil integrated development environment. Through the production of the project, I achieved all the set goals. The program can be used with any lens, it focuses the image regardless of the picture color and its operation should be unaffected by any external factors such as weather conditions. The program is now in use by the DATCON company as part of their sales offering.

• Keywords: Lens, Autofocus, PID controller, PWM, Brenner gradient