Effective Noise Reduction Methods for Rear-View Monitoring Devices Based on Microprocessors

Abstract.

This paper suggests effective noise reduction methods for the neckband devices which are able to monitor the rear-view images. This device helps the user to keep an eye on the rear areas which he/she cannot see in normal ways without turning back. In order to be used as a wearable device, the neckband device should have some particular properties such as small sizes, light weight, and low power consumption. When users walk or move, the neckband device also moves, which causes many noise effects to the system such as illumination changes and especially severe camera motion. We propose effective noise reduction methods to deal with these noise effects. The experiments demonstrate that the proposed methods have good performance in rear-view monitoring under the practical difficulties that arise due to users’ movement such as illumination changes and out of focus problems.