A localization approach based on adaptive threshold segmentation for mobile robots

To adapt to the illumination and gray variance changed largely in room during autonomous navigation for mobile robot, an adaptive image processing approach is proposed based on adaptive threshold segment. The pixel coordinates in video are picked up in real time and fitted into a line. According to the included angle between the fitted line and the horizontal line, the robot pose is emendated to move along with the horizontal line. An improved localization algorithm for mobile robot is proposed to track robot pose in real time according to the information from CCD and odometry sensors based on the Extended Kalman theory. The experiments have denoted that the adaptive image segment algorithm detects the object effectively, and the improved pose-tracking algorithm reduces localization errors, improves the localization accuracy to a great extent.