Image Processing Based on Inter-frame Color Feature Matching for Visible Light Positioning

The indoor imaging visible light positioning (VLP) technology based on the image sensor (IS) utilizes existing indoor lighting infrastructure to provide high-precision indoor positioning services. However, due to the high complexity of traditional image processing algorithms, existing VLP systems are unable to provide real-time positioning services. In response to the above issue, we propose an image processing mechanism based on the inter-frame color feature matching method, which includes the equidistant sampling processing, low complexity decoding algorithm, and color feature matching-based LED region of interest (ROI) determination method. The results show that the system average positioning delay is 80.2ms, indicating better system real-time performance.

I. Introduction

In recent years, with the rapid development of Light Emitting Diode (LED), visible light positioning (VLP) technology has gradually become a powerful alternative to wireless positioning technology^[1]-[2]. According to the receiver types, the VLP system can be divided into the non-imaging based VLP system and imaging based VLP system^[3]-[4]. Compared the with non-imaging based VLP system, the imaging based VLP system can separate light signals by spatial projection, which is more suitable for weak light and complex environments. Furthermore, the imaging based VLP system has no limit on the number of captured LEDs and can provide users with more accurate positioning services^[5]-[6]. However, the imaging based VLP system requires multiple steps in the image processing process, such as detection, segmentation, decoding, etc. These steps require a large amount of computation on the pixels, which increases the system positioning latency and has an obvious improvement on the real-time performance.