움직인 피사체를 선택적으로 표현하는 파노라마 이미지 스티칭

초록

In this paper, we propose an panorama image stitching method that selective representation of moving objects in images. The moving objects are caught in stitched images, and it leads to weird result such as ghost. The proposed method can prevent that result by detecting the object and its selective representation. The proposed method uses Gaussian blurred image for robust to noise and speeded-up robust features (SURF) and random sample consensus (RANSAC) are also used for image stitching. In the experiments, we verify the proposed method can detect a moving objects area and through out this area, the moving object can be rightly seen in the image. 행사명: 2017년도 대한전자공학회 정기총회 및 추계학술대회 행사 기간: 2017년 11월 24일 ~ 25일 행사 장소: 송도 컨벤시아 주관 기관: 대한전자공학회