Reducing Grayscale Value Fluctuation for Mobile-Phone Camera Based VLC System

Visible light communication (VLC) technology based on mobile-phone camera has attracted more and more attention. Using the rolling-shutter effect of a complementary metal-oxide-semiconductor (CMOS) camera, the VLC system can achieve a higher data rate than CMOS camera frame rate. However, the blooming effect of the CMOS camera and light noise can cause grayscale value fluctuation, which can cause data logic not being correctly defined by a decision threshold, thereby affecting system performance. In this paper, for the first time, we propose an exponent transform (ET) algorithm to decrease grayscale value fluctuation, thus improving the VLC system performance. Compared with our previously proposed moving exponent average algorithm, the ET algorithm has advantages in computational complexity and bit error rate performance. Experimental results show that with the ET algorithm, over 50 cm (150 cm) transmission distance at a data rate of 5 Kb/s (2.5 Kb/s) can be achieved under the 7% forward error correction limit (BER = 3.8 × 10
^-3
) for the mobile-phone camera based VLC system.