Oluwatobi-coder/GIS-Crop-Coefficient-Estimation

This project implements a Remote Sensing & Machine Learning pipeline to estimate spatial Crop Coefficients ($K_c$) in Bacita, Nigeria. By integrating Google Earth Engine (GEE) with local field data, the system models the relationship between spectral vegetation indices (like NDVI) and crop water requirements. # GIS-Based Crop Coefficient ($K_c$) Estimation for Precision Agriculture This project implements a **Remote Sensing & Machine Learning** pipeline to estimate spatial Crop Coefficients ($K_c$) in Bacita, Nigeria. By integrating Google Earth Engine (GEE) with local field data, the system models the relationship between spectral vegetation indices (like NDVI) and crop water requirements. ## 📌 Project Overview Accurate estimation of evapotranspiration is vital for efficient irrigation scheduling. Traditional single-coefficient methods often fail to capture spatial variability in large fields. This project leverages satellite imagery and **Linear Regression** to map $K_c$ distribution, validating the results against ground-truth field data (`bacita_true_field_data.csv`). ## 🛠️ Tech Stack * **Geospatial Engine:** Google Earth Engine (EE) API * **Interactive Mapping:** Geemap * **Data Manipulation:** Pandas, NumPy * **Machine Learning:** Scikit-Learn (Linear Regression) * **Visualization:** Matplotlib, Seaborn ## ⚙️ Key Features * **Cloud-Native Processing:** Fetches and processes satellite imagery (Sentinel-2/Landsat) using GEE without local download. * **Spectral Analysis:** Automatically calculates vegetation indices (NDVI, SAVI) to serve as predictors for $K_c$. * **ML Calibration:** Uses `sklearn` to regress field-observed data against satellite indices to derive a localized predictive model. * **Interactive Exports:** Generates static heatmaps and interactive HTML maps (`Bacita_Final_Project.html`) for field deployment. ## 📂 Methodology & Data The workflow follows a standard remote sensing pipeline: * **Input Data:** * **Satellite:** Multi-spectral imagery from Earth Engine. * **Ground Truth:** `data/bacita_true_field_data.csv` containing observed field measurements. * **Processing:** Cloud masking -> Index Calculation -> Spatial Resampling. * **Modeling:** $K_c = \alpha \times NDVI + \beta$ (calibrated via Linear Regression). ## 📁 Repository Stru …