Advancing Cassava Age Estimation in Precision Agriculture: Strategic Application of the BRAH Algorithm

Abstract

Cassava crop age estimation is crucial for optimizing irrigation, fertilization, and pest management, which are key components of precision agriculture. Accurate knowledge of crop age allows for effective resource application, minimizing environmental impact and enhancing yield predictions. The Bare Land Referenced Algorithm from Hyper-Temporal Data (BRAH) is used for bare land classification and cassava crop age estimation, but it traditionally requires manual NDVI thresholding, which is challenging with large datasets. To address this limitation, we propose automating the thresholding process using Otsu’s method and enhancing the image contrast with histogram equalization. This study applies these enhancements to the BRAH algorithm for bare land classification and cassava crop age estimation in Ratchaburi, Thailand, utilizing a dataset of 604 Landsat satellite images from 1987 to 2024. Our research demonstrates the accuracy and practicality of the BRAH algorithm, with Otsu’s method providing 94% accuracy in detecting the bare land validation locations with an average deviation of 8.78 days between the acquisition date and the validated date. This approach facilitates precise agricultural planning and management, promoting sustainable farming practices and supporting several Sustainable Development Goals (SDGs).