نوع مقاله : مقاله پژوهشی
نویسندگان
1 گروه علوم و مهندسی آب، دانشکده کشاورزی و منابع طبیعی، دانشگاه بین المللی امام خمینی (ره)، قزوین، ایران.
2 گروه علوم و مهندسی آب، دانشکده کشاورزی و منابع طبیعی، دانشگاه بین المللی امام خمینی(ره)، قزوین، ایران
3 بخش تحقیقات فنی و مهندسی کشاورزی، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی خوزستان، سازمان تحقیقات، آموزش و ترویج کشاورزی، اهواز،
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Accurate Manitoring of date-palm plantations and reliable estimation of cultivated area are essential for water-resource management, agricultural planning and regional policy making in arid and semi-arid regions, specially in south-west of Iran. This research utilized multi-temporal, Sentinel-2, and radar, Sentinel-1 imageries and supervised classification based on random forest, to present accurate and applicable map of date-palms plantations in Maroon-Jarahi sub-basin.
To address this, Sentinel-2 optical and Sentinel-1 radar time series were derieved and processed, then together with Spectral, Radar, thermal indices and collected field reference samples during the 2022-2024 agricultural seasons were applied to training process of random forest model. Classification was run based on 36 inputs including, main bands, Spectral, Radar, and thermal indices. Accuracy assessment of the Model according to independent sample test displayed that The Random Forest classifier yielded very high performance due to, overall accuracy= 95%, kappa coefficient = 0.94, and precision, recall and F1-score were approximately 0.99. furthermore, mapped date-palm area closely matched reference inventories, the actual area was 26197.55 ha while the mapped area was 21764.55 ha for the Random Forest result, thus approximately +1% over-estimation was derieved. The results showed that, The integration of multi-temporal and thermal indicators within a Random Forest classification framework provides an efficient, reliable and scalable method to map date-palm plantations in arid and semi-arid basins.Practically, the proposed workflow can be a validated approach for water resource management, Orchards’ health monitoring and agricultural planning in reginal and national scales.
کلیدواژهها [English]