Using Enhanced-Color Mapping Algorithm for Object Boundary Segmentation

摘要

Image segmentation and object extraction boundaries in thermal images are fundamental approach in digital image processing in obtaining pattern recognitions and alike. A method like the Otsu, improves the image segmentation effect from the thresholding process of separated objects in the background image. This proposed algorithm integrates the three significant outputs: the extracted high thermal threshold value from the original thermal image, and the two outputs from the inverted Otsu method which processed by Canny edge detection and color mapping. Also, this paper is synthesized with Xilinx Zync 7000 ZED ZC702. The experimental results show that the separated objects can easily be detected from its background and shows a significant distinguishable features in terms of edge and color far better against the conventional thermal image.