Penggunaan Algorithm Watershed Untuk Segmentasi Penutupan Tajuk Di Hutan Lahan Kering

Abstract

Makalah ini menjelaskan penggunaan algoritma watershed untuk mendeteksi penutupan tajuk dihutan lahan kering. Tujuan penelitian adalah untuk menentukan parameter terbaik dari algoritma watershed segmentasi untuk memperoleh informasi penutupan tajuk dari citra resolusi tinggi dan sangat tinggi yang difilter dan tidak difilter. Performa terbaik dari masing-masing kombinasi parameter nilai toleransi (T), nilai mean (M), dan nilai varians (V), yang ditulis sebagai C:[T]-[M]-[V], ditentukan berdasarkan tingkat akurasi. Penelitian ini menggunakan citra Pleiades-1B dan SPOT-6 sebagai data digital primer. Hasil penelitian menunjukkan bahwa citra Pleiades-1B terfilter low-pass menunjukkan performa terbaik dengan kombinasi parameter C6-MF:[10]-[0.7]- [0.3], memiliki akurasi overall (OA) 90,40% dan Kappa akurasi (KA) sebesar 83,2%. Sedangkan pada citra Spot-6 terfilter low-pass menunjukan kombinasi parameter C7-MF:[10]-[0.8]-[0.2], memiliki akurasi OA 90,6% dan KA 65,4%. Penelitian ini menyimpulkan bahwa citra hasil filter dengan low-pass filter selalu memberikan hasil yang lebih akurat dibandingkan data aslinya (tanpa filter), baik untuk citra Pleiades-1B maupun SPOT-6. Resolusi spasial sangat tinggi memberikan akurasi yang lebih baik daripada resolusi spasial tinggi

The sight of increasingly high-resolution images in the forestry sector makes it increasingly essential to find methods for detecting and identifying forest stand parameters that are more accurate. It is no longer possible to rely solely on a direct pixel-based classification approach to detect one of the stand variables in high-resolution images, namely tree canopy closure. The pixel-based approach is wildly inaccurate when applied to images with very high spatial resolution. Segmentation will enable classify objects into superpixel classes (segments) before further classification. This paper uses a watershed algorithm to detect canopy cover in dryland forests. The study at to determine the best parameters of the watershed segmentation algorithm to obtain information on crown closure from filtered and unfiltered high and very high-resolution images. The best performance of each parameter combination of tolerance value (T), mean value (M), and variance value (V), which is written as C:[T]-[M]-[V], is determined based on the level of accuracy. This study uses Pleiades-1B and SPOT-6 images as primary digital data. The results showed that the low-pass filtered Pleiades-1B image showed the best performance with a combination of parameters C6-MF:[10]-[0.7]-[0.3], had an overall accuracy (OA) of 91.0% and an accuracy Kappa (KA) by 83.2%. While the low-pass filtered Spot-6 image shows the combination of parameters C7-MF:[10]-[0.8]-[0.2], which has an accuracy of 90.6% OA and 65.4% KA. This study concludes that the filtered image with a low-pass filter always gives more accurate results than the original data (without filter), both for Pleiades-1B and SPOT-6 images. The very high spatial resolution provides better accuracy than the high spatial resolution