IDENTIFICATION OF SMALLHOLDERS' OIL PALM PLANTATIONS USING ALOS DATA (A CASE STUDY IN LAMPUNG PROVINCE, INDONESIA)

摘要

Oil palm is one of the most profitable crops for developing countries in tropical area. The expansion of oil palm cultivation has been rapidly increased to meet growing demands of palm oil, but it is also considered as a great threat for tropical ecosystems. Sustainable monitoring, management, and environmental effect assessment of this expansion is important to lower its contribution to biodiversity loss and climate change, while also fostering vigorous national economic development. As one of the efforts, accurate identification of all management types of oil palm plantations is necessary. Remote sensing technology is known to have high potential to identify the spatial distribution of any land cover types. However, the identification of smallholders' oil palm plantations is a great challenge since its sparse distribution in small patches is still practically hard to be identified using currently available remote sensing data. The present study aimed to assess the ability of textural analysis to detect oil palms and to identify the smallholders' oil palm plantations in Mesuji District, Lampung Province, Indonesia by using both of microwave and optical images of the Advanced Land Observing Satellite (ALOS). Texture analysis using eight texture features in six types moving windows was examined to extract the triangular planting pattern from ALOS Phased Array type L-Band Synthetic Aperture Radar (PALSAR) dual polarimetry data. The results showed mean and variance features in 11 x 11 moving window as the best combination for identifying oil palms. An improvement in the classification accuracy was obtained through the integration of the texture analysis result with ALOS Advanced Visible and Near Infrared Radiometer type-2 (AVNIR-2) image. This method also revealed a great possibility to differentiate between the young and mature oil palms.