Rice leaf chlorophyll content estimation with different crop coverages based on Sentinel-2

Abstract

Chlorophyll content is an important index for evaluating the health and productivity of crops, and environmental stress on them. The real-time, rapid, and accurate acquisition of chlorophyll content plays a key role in crop growth monitoring. Remote sensing can quickly obtain chlorophyll content in regional and global scale, but how to eliminate the interference of soil background in estimation is a major challenge. The statistical analysis method based on empirical/semi-empirical model is simpler, faster and easier to implement than that based on radiative transfer mechanism model. The statistical analysis method can eliminate the influence of environmental background by looking for a Vegetation index (VI) that is sensitive to the chlorophyll content but not to soil background to a certain extent. However, the accuracy of this method is low in fields with different degrees of crop coverage. Additionally, the special soil background characteristics of rice fields make the method a doubtful tool to estimate the Chlorophyll content in rice coverage. To improve the accuracy of the rice leaf chlorophyll content (LCC) estimation model, we here propose a new model based on crop coverage. We analyzed remote sensing images of the Sentinel-2 over rice-planting areas of Qian Gorlos County in Jilin Province of China. We divided the study area into three regions based on high-, medium-, and low-rice canopy coverage. Rice LCC in each region was estimated by identifying remote sensing vegetation indices that are sensitive to chlorophyll in different rice canopy coverages. Compared to the estimated results without considering the crop coverage, our model achieves a higher accuracy. In addition, we applied the model in the region of Northeast China in 2023 to verify its strong generalisability and robustness. Our study provides a reference for rapidly and non-destructively obtaining rice LCC based on Sentinel-2 images. However, the applicability of our model to other crops will be verified in the future.