Optimal Color Composition Method for Generating High-Quality Daily Photographic Time Series From PhenoCam

Abstract

Phenology camera (PhenoCam) data and the derived green chromatic coordinate (GCC) time series are commonly used to track seasonal changes in canopy greenness. However, the GCC time series is noisy because color distortion commonly exists in the captured photographs owing to the varying illumination conditions and nonlinear response of the consumer-grade camera to the incoming light. Hence, we proposed an optimal color composition (OCC) method to generate high-quality daily photographic time series by compositing multiple photographs captured in a single day. First, the optimal acquisition time with good illumination conditions and correct exposure settings is determined for each pixel throughout the day based on a comprehensive color index, combining the brightness and saturation. A virtual photograph consisting of the selected digital numbers acquired at the optimal time is then composited for each day. Finally, the daily GCC time series is calculated based on the virtual photographs. By testing the photographs of six forest sites, the proposed method was compared with the commonly used 90th percentile (Per90) filter. The results show that the daily photographs composited using the OCC method were more homogeneous with less shaded areas compared to those selected by the Per90 filter, and the corresponding GCC time series derived from the OCC method is more stable and less influenced by varying atmospheric conditions and solar angles than the Per90 filter. These results indicate that the OCC method can generate high-quality daily photographic time series with the potential to better indicate seasonal color changes in the forest canopy.