Modified Otsu's method for indoor mobile robot tracking system
Sewon Lee, Jin Won Jang, Kwang‐Ryul Baek, Heungbo Shim · 2014
In vision-based tracking system, thresholding is one of the most important steps in image pre-processing. Thresholding algorithm has a strong influence on both accuracy and performance in object tracking. Thresholding algorithms are classified as global thresholding or local thresholding. In general, the computing power required for local thresholding algorithm is more than ten times that of global thresholding algorithm, so global thresholding algorithm is suitable for a real-time application. The Otsu's method is the most famous global thresholding algorithm, however, it misclassifies object as background in some cases. To reduce the misclassification problems, we apply the modified Otsu's method for indoor mobile robot tracking system. Experimental results show that applied algorithm improves the performance of thresholding results.