دانشگاه بیرجند پژوهش‌های زعفران 2345-3869 13 2 2025 09 23 Mapping Saffron Cultivation Areas Using Landsat 8-9 Time-Series and a Pixel-Based SVM Classification Method Mapping Saffron Cultivation Areas Using Landsat 8-9 Time-Series and a Pixel-Based SVM Classification Method 353 367 3957 10.22077/jsr.2026.11019.1301 FA میلاد جانعلی پور پژوهشگاه هوافضا، وزارت علوم تحقیقات و فناوری، تهران، ایران. نادیا عباس زاده طهرانی پژوهشگاه هوافضا، وزارت علوم تحقیقات و فناوری، تهران، ایران. محمد جعفری سنجش از دور و سیستم اطلاعات جغرافیایی، دانشگاه آزاد اسلامی، تهران، ایران. Journal Article 2026 02 19 Remote sensing provides an efficient and cost-effective way to estimate the cultivation area of agricultural crops. In this study, the cultivation area of saffron in Roshtkhar city was estimated using Landsat 8 and 9 time-series imagery and a pixel-based Support Vector Machine (SVM) classification method. The analysis, conducted for the Persian calendar year 1402 (2023 AD), identified approximately 450 hectares of saffron cultivation. An accuracy assessment indicated that the proposed method performed robustly. Based on test samples, classification of Landsat 8 imagery yielded an overall accuracy of approximately 92% and a user's accuracy of 95% for the saffron class. Results from Landsat 9 imagery were similarly high, with an overall accuracy of roughly 91% and a saffron class accuracy ranging from 90% to 93%. Consequently, while Landsat 8 demonstrated slightly superior performance in identifying saffron lands, both sensors proved highly effective for this application, confirming the suitability of the proposed method for distinguishing saffron from other land cover classes. Remote sensing provides an efficient and cost-effective way to estimate the cultivation area of agricultural crops. In this study, the cultivation area of saffron in Roshtkhar city was estimated using Landsat 8 and 9 time-series imagery and a pixel-based Support Vector Machine (SVM) classification method. The analysis, conducted for the Persian calendar year 1402 (2023 AD), identified approximately 450 hectares of saffron cultivation. An accuracy assessment indicated that the proposed method performed robustly. Based on test samples, classification of Landsat 8 imagery yielded an overall accuracy of approximately 92% and a user's accuracy of 95% for the saffron class. Results from Landsat 9 imagery were similarly high, with an overall accuracy of roughly 91% and a saffron class accuracy ranging from 90% to 93%. Consequently, while Landsat 8 demonstrated slightly superior performance in identifying saffron lands, both sensors proved highly effective for this application, confirming the suitability of the proposed method for distinguishing saffron from other land cover classes. https://jsr.birjand.ac.ir/article_3957_16ce9310310c5e5005400bf14c813930.pdf