Optimal protective region selection algorithm for patrol robots

This paper presents an optimal protection region selection algorithm based on color and texture. It employs a fuzzy approach to evaluate the similarity of color, in which morphological open operation and close operation are used to remove the noise in images. Moreover, the protection area is evaluated from the aspect of texture similarity and color similarity. With this approach, in patrol missions, robots could choose the best protective region-a region in which they are most unlikely to be recognized, so as to achieve the goal of hiding-themselves when approaching their targets. Two robots are applied to test the proposed algorithm in several different environments, namely a railway station, a warehouse and a campus. Test results show that this approach could select protective regions for patrol robots automatically, however, the method proposed can be influenced by the illumination and the view angle, needs further improvement.