Development and performance analysis of Canny and Hough transform based pole detection algorithm as mobility aid for visually impaired person on SEE-S (Smart Environment Explorer-Stick)

This study investigated approach to improve the SEE-S (Smart Environment Explorer-Stick) function in detecting obstacle for Visually Impaired Person (VIP). The study proposed enhanced algorithm to detect the presence of poles. This algorithm uses a combination of method of distance calculation and pairing search of vertical lines obtained from optimized Canny detection. The position of the vertical lines was found using Houghlines function. Our field performance evaluation tests indicated that the proposed algorithm works satisfactorily. Two dominant variables that significantly affect the ability of SEE-S in detecting poles were identified. They were the location of the test and the colour of poles. Two pole colours are used, namely black and silver. At indoor the algorithm performance was worse than at outdoor by 0.085 times. The use of silver pole improved the performance of the algorithm by 11.757 times.