Year: 2026
Type: article
Abstract: The tea plant (Camellia Sinensis), as the world's most popular non-alcoholic beverage, underscores the importance of precise and timely spatial data for industry insights and sustainability. Yet, the accurate delineation of spectrally similar vegetation types, notably tea trees, continues to elude conventional methods. This study innovatively integrates Sentinel-1 radar data with Sentinel-2 imagery to effectively overcome optical observation limitations imposed by Yunnan's cloudy climate, particularly during non-growing seasons (e.g., February). By systematically analyzing annual phenological dynamics, we quantitatively identified April as the optimal temporal window for discriminating tea trees from spectrally similar vegetation, such as rubber and natural forests. Furthermore, optimizing the model through a feature selection process to eliminate redundant features significantly enhanced the overall classification accuracy from 87.1% to 89.1%. This study could assistant monitoring the crop dynamics and timely respond to cultivation activity.
Source: Frontiers in Plant Science
Authors: Nan Cong, Rongrong Zhao, Yuxin Qiu, Chuang Zhao
Language: English
Subfield: Ecology
Field: Environmental Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7161168419