Oil Palm Counting and Age Estimation from WorldView-3 Imagery and LiDAR Data Using an Integrated OBIA Height Model and Regression Analysis

• Main Authors: Hossein Mojaddadi Rizeei, Helmi Z. M. Shafri, Mohamed Ali Mohamoud, Biswajeet Pradhan, Bahareh Kalantar
• Format: Article
• Language: English
• Published: Hindawi Limited 2018-01-01
• Series: Journal of Sensors
• Online Access: http://dx.doi.org/10.1155/2018/2536327

The current study proposes a new method for oil palm age estimation and counting. A support vector machine algorithm (SVM) of object-based image analysis (OBIA) was implemented for oil palm counting. It was integrated with height model and multiregression methods to accurately estimate the age of trees based on their heights in five different plantation blocks. Multiregression and multi-kernel size models were examined over five different oil palm plantation blocks to achieve the most optimized model for age estimation. The sensitivity analysis was conducted on four SVM kernel types (sigmoid (SIG), linear (LN), radial basis function (RBF), and polynomial (PL)) with associated parameters (threshold values, gamma γ, and penalty factor (c)) to obtain the optimal OBIA classification approaches for each plantation block. Very high-resolution imageries of WorldView-3 (WV-3) and light detection and range (LiDAR) were used for oil palm detection and age assessment. The results of oil palm detection had an overall accuracy of 98.27%, 99.48%, 99.28%, 99.49%, and 97.49% for blocks A, B, C, D, and E, respectively. Moreover, the accuracy of age estimation analysis showed 90.1% for 3-year-old, 87.9% for 4-year-old, 88.0% for 6-year-old, 87.6% for 8-year-old, 79.1% for 9-year-old, and 76.8% for 22-year-old trees. Overall, the study revealed that remote sensing techniques can be useful to monitor and detect oil palm plantation for sustainable agricultural management.