Elevator‘s External Button Recognition and Detection for Vision-based System

Abstract

Recently, autonomous transporter offers the assistance and delivery for user but they are only focusing on single floor environment. To widen up fields of robotic, they teach robot to use an elevator because elevator provides an essential means of faster movement across level. However, most of the mobile service robot failed to detect elevator’s position due to the complex background and reflection on the elevator door and button panel itself. This paper presents a new strategy for recognition method to detect elevator by detecting their external button efficiently. Sobel is use as edge detection operator to find the estimated absolute gradient magnitude at each point in an input grayscale image. Then, but we enhanced the technique by combining it with wiener filter to reduce the amount of noise present in a signal by comparing the signal with an estimation of the desired noiseless signal. This filter helps to eliminate the reflection image on elevator’s button panel before it can be converted to black and white image (binarization). The process followed by some morphological and structuring elements process. Tests have been done and the results shown that elevator’s external button can be recognized and detected by those entire framework.