Estimation of Rice Plant Coverage Using Sentinel-2 Based on UAV-Observed Data

Yuki Sato,Takeshi Tsuji,Masayuki Matsuoka

DOI: https://doi.org/10.3390/rs16091628

IF: 5

2024-05-02

Remote Sensing

Abstract:Vegetation coverage is a crucial parameter in agriculture, as it offers essential insight into crop growth and health conditions. The spatial resolution of spaceborne sensors is limited, hindering the precise measurement of vegetation coverage. Consequently, fine-resolution ground observation data are indispensable for establishing correlations between remotely sensed reflectance and plant coverage. We estimated rice plant coverage per pixel using time-series Sentinel-2 Multispectral Instrument (MSI) data, enabling the monitoring of rice growth conditions over a wide area. Coverage was calculated using unmanned aerial vehicle (UAV) data with a spatial resolution of 3 cm with the spectral unmixing method. Coverage maps were generated every 2–3 weeks throughout the rice-growing season. Subsequently, crop growth was estimated at 10 m resolution through multiple linear regression utilizing Sentinel-2 MSI reflectance data and coverage maps. In this process, a geometric registration of MSI and UAV data was conducted to improve their spatial agreement. The coefficients of determination (R2) of the multiple linear regression models were 0.92 and 0.94 for the Level-1C and Level-2A products of Sentinel-2 MSI, respectively. The root mean square errors of estimated rice plant coverage were 10.77% and 9.34%, respectively. This study highlights the promise of satellite time-series models for accurate estimation of rice plant coverage.

environmental sciences,geosciences, multidisciplinary,imaging science & photographic technology,remote sensing

What problem does this paper attempt to address?

The main objective of this paper is to develop a method to estimate the coverage of rice in paddy fields (i.e., the proportion of the ground covered by rice plants) using Sentinel-2 satellite data combined with data observed by Unmanned Aerial Vehicles (UAV). Specifically, the research team aims to monitor the growth status of rice and improve estimation accuracy through this method.

To achieve this goal, the researchers adopted the following steps:

1. **Data Collection**:
- High-resolution images were collected at different time points using a DJI P4 Multispectral UAV equipped with a multispectral camera.
- Image data with high temporal and spatial resolution were obtained using the multispectral instrument (MSI) of the Sentinel-2 satellite.

2. **Preprocessing**:
- Orthorectification and radiometric correction were performed on the UAV images to ensure image quality.
- Geometric registration was conducted to improve spatial consistency between UAV images and Sentinel-2 images.
- Rice field pixels were extracted to focus on the target area.

3. **Rice Coverage Calculation**:
- Linear spectral unmixing techniques were applied to calculate the area proportion of rice and water in each mixed pixel.
- Constrained least squares method was used to determine rice coverage.

4. **Model Building**:
- A model was established through multiple linear regression analysis based on the rice coverage maps obtained from UAV data.
- The model was constructed using Sentinel-2 MSI reflectance data and rice coverage maps to assess rice growth conditions.

5. **Validation and Evaluation**:
- The impact of different Sentinel-2 products (Level-1C and Level-2A) on model accuracy was compared.
- The applicability of the model to different rice varieties was evaluated.

The research results indicate that by combining UAV observation data with Sentinel-2 satellite data, rice coverage can be effectively estimated, thereby supporting the monitoring of rice growth conditions. This method can improve estimation accuracy and contribute to the intelligent development of agricultural management.