| Title |
Development of an Auto-focusing Algorithm Based on Random Forest |
| Authors |
윤경호(Kyeong-Ho Yoon) ; 이대종(Dae-Jong Lee) ; 전명근(Myung-Genu Chun) |
| DOI |
https://doi.org/10.5370/KIEEP.2020.69.1.55 |
| Keywords |
Auto-focusing; Random forest; Focusing value; Machine vision |
| Abstract |
Auto-focusing is essential for the detection of minute defects in inspection equipments. We propose an optical system that can autofocus on a glass substrate and an algorithm that can adjust the focus quickly and accurately. The optical system uses Koehler illumination to measure the contrast of the field of view aperture, which is effective for plain glass. The auto-focusing algorithm using random forest is learned by setting the target value for the input image with the estimated focal length. When a new image is given, the focal length estimate is calculated. Various experiments have been performed on focus measurement using images acquired from various locations and show better results than others. |