This paper focuses on a road-to-vehicle visible light communication (VLC) system using LED traffic lights and high-speed cameras on cars. So as to improve the transmission reliability, an error correcting code like a low-density parity-check (LDPC) is applied to the road-to-vehicle VLC system. The decoder of the LDPC code needs channel information in order to maximize decoding performance. Usually, it is calculated from a signal-to-noise ratio (SNR) or bit error rate (BER) of a channel, where noise is a dominant factor of the channel characteristic. In the road-to-vehicle VLC system, not noise but interference among LEDs decides the channel characteristic. In this paper, we propose a channel information estimation method for the road-to-vehicle VLC system. The proposed method estimates a signal-to-interference ratio (SIR) from the receiving image, and calculates the channel information using the SIR. The proposed method is evaluated by experiment. As a result, the proposed method can achieve almost the same performance with the optimum case, and then maximize the decoding performance.