การตรวจหาจอภาพคอมพิวเตอร์ในภาพนิ่งซึ่งจับมาจากกล้องวีดิทัศน์ / สมจิตต์ อัครวรมาศ = Computer monitor detection in still images captured a video camera / Somchit Akaraworamas
The objective of this research present an approach to the detection of monitors in still images captured from video camera. The monitors are analog image sequence use PAL standard at resolution 768x576 and 16 bit colors depth, transfer to digital images in AVI format and decrease color bit depth to 8 bit grayscale. Then, the quality of images is improved by "Average Image" method and histogram equalization. After that, edges are detected and edges discontinue were corrected by Hough & De-Hough Transform method. Finally, object edges and the boundary of template model are compared, if the edge of object and template model are matched, then it is the same object. The research was tested on 200 images sequence captured from computer laboratories and each images sequence has 10 frames picture. This research involves studies of the procedure and limitations to apply the images captured from video camera, and method that is solving problems of images with blur images, dark or light images and objects that are overlapped. From 200 images sequence taken at different Computer Engineering laboratories under fluorescent light, the result of the research from proposed procedure can detected about 90 percent of computer monitors. The resent for undetected monitors are bad quality images and the images where the monitors are overlapped
