A fixed-threshold method for estimating fractional vegetation cover of maize under different levels of water stress

Authors: NIU, YAXIAO - Northwest A&f University; Zhang, Huihui; WENTING, HAN - Northwest A&f University; ZHANG, LIYUAN - Northwest A&f University

Submitted to: Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/2/2021
Publication Date: 3/7/2021
Citation: Niu, Y., Zhang, H., Wenting, H., Zhang, L. 2021. A fixed-threshold method for estimating fractional vegetation cover of maize under different levels of water stress. Remote Sensing. 13(5). Article e1009. https://doi.org/10.3390/rs13051009.
DOI: https://doi.org/10.3390/rs13051009

Interpretive Summary: With the availability of commercial high-resolution RGB cameras, we are able to capture crop information easily. However, accurate estimation of fractional vegetation cover (FVC) from digital images taken by these cameras needs be improved to monitor crop growth status, especially when plants are under water stress. Two classic threshold-based methods, the intersection method (T1) and the equal misclassification probability method (T2), have been widely applied to RGB images. However, high coverage and severe water stress of crops in the field make it difficult to extract FVC accurately. To solve this problem, this study proposed a fixed-threshold method based on the statistical analysis of thresholds obtained from the two classic threshold approaches. Results showed that, for images with high reference FVC, the proposed fixed-threshold method solved the problem of underestimation from T1 (12.8%) or T2 (14.1%) methods. Compared to T1 and T2 methods, for images taken in severe water stress plots, the mean estimation error of FVC obtained by the fixed-threshold method was decreased by 4.3% and 19.3%, respectively. Overall, the FVC estimates from the proposed fixed-threshold method are robust, accurate and efficient, with R2 of 0.99 and RMSE of 0.02.

Technical Abstract: Accurate estimation of fractional vegetation cover (FVC) from digital images taken by commercially available cameras is of great significance to monitor the vegetation growth status, especially when plants are under water stress. Two classic threshold-based methods, the intersection method (T1 method) and the equal misclassification probability method (T2 method), have been widely applied to RGB images with high spatial resolution. However, the high coverage and severe water stress of crops in the field make it difficult to extract FVC stably and accurately. To solve this problem, this paper proposed a fixed-threshold method based on the statistical analysis of thresholds obtained from the two classic threshold approaches. Firstly, the Gaussian mixture model (GMM), including the distributions of green vegetation and backgrounds, was fitted on four color features: excessive green index, H channel of HSV color space, a* channel of the CIE L*a*b* color space, and the brightness-enhanced a* channel (denoted as a*_I). Secondly, thresholds discriminating green vegetation were calculated by applying T1 and T2 method to the GMM of each color feature. Thirdly, based on the statistical analysis of the thresholds with better performance between T1 and T2, the fixed-threshold method which includes an initial threshold and an adjusted threshold was proposed. Finally, the fixed-threshold method was applied to the optimal color feature a*_I to estimate FVC, and compared with the two classic approaches. Results showed that, for some images with high reference FVC, FVC were serious underestimated by 0.128 and 0.141 when using T1 and T2 method, respectively, which was eliminated by the proposed fixed-threshold method. Compared to T1 and T2 methods, for images taken in serious water stress plots, the mean estimation error of FVC obtained by the fixed-threshold method was decreased by 0.043 and 0.193, respectively. Overall, the FVC estimations of the proposed fixed-threshold method has the advantages of robustness, accuracy and high efficiency, with R2 of 0.99 and RMSE of 0.02.