The Threshold Optimization of Track Image Under Strong Sunlight

The threshold optimization of track image plays an important role in intelligent car competition. The traditional binary imaging method is not ideal under strong light. After using mean filter to convolve specific image, we explore first-order difference to extract the trend of histogram, and obtain all peaks and troughs that satisfy corresponding condition in the histogram. As a result, the position of rightmost can be selected as the optimal threshold. The threshold value obtained by our proposed method can effectively extract the track part of the image, thus providing an accurate guarantee for the direction control of intelligent car.