html
This work aims at proving a tool for robot localization purposes in different environments based on the recognition of visual features. Features include edges and other distinctive shapes. Visual data was acquired form a single camera. The data set was utilized for the determination of the robot’s environment. Two methods were used to perform localization. The first method is based on points of interest, whereas, the second method uses stable region information from the image. An algorithm for the recognition of stable areas in the image is proposed in this manuscript. Results show a stable recognition of image features for environment recognition and localization purposes.
