Hozuma Sekine

Application of hyperspectral data for assessing peatland forest condition with spectral and texture classification

Peatland in tropical region is a major CO2 emission source because of peat decomposition and forest fire by human induced activities. Remote sensing is effective tool to monitor environmental condition of peatland and forest ecosystem in peatland. A pixel-based approach is one of the most attractive choices for forest type classification or biomass prediction. The traditional method, however, is not sufficient for using spatial information. The spatial information, such as image texture, is an important factor for identifying objects or types, because a pixel is not independent of its neighbors and its dependence can be useful for classification and biomass prediction in forest regions. In this paper, we used combined data of spectral and spatial information from hyperspectral data (Hymap) to develop a more accurate classification or biomass prediction model. The spatial information was texture data by using Grey Level Co-occurrence Matrix (GLCM) texture measures. Sparse discrimination analysis (SDA) was applied for the classification model, and LASSO regression was applied for the biomass prediction model. The results were compared to find out how the spatial information enhances the classification and biomass prediction. According to the accuracy assessment, both classification and biomass prediction model derived from the combined data performed high accuracy.

Validation of BiPLS for improving yield estimation of rice paddy from hyperspectral data in West Java, Indonesia

Rice is one of the important agricultural crops and constitutes major staple food for Asian countries, especially in Indonesia. Monitoring and management of paddy fields are one of key factors for ensuring national food security, and remote sensing technology and its data, especially hyperspectral data, are expected to be a highly effective solution. The objective of this paper is to use airborne hyperspectral data and yield data to develop and validate a high performance prediction model based on a statistical technique for yield of rice paddy.

A methodology of forest monitoring from hyperspectral images with sparse regularization

This paper presents a methodology to extract information on existing conditions of a forest from hyperspectral images and SAR images for the forest management. To overcome the difficulties in hyperspectral image analysis such as optimal band selection and model overfitting, a machine learning technique called sparse regularization was adopted. Experimental results show the effectiveness of this approach.

Sensing/Monitoring Networks on Carbon Balance and Biodiversity in Tropical Peatland

The Earth’s remaining tropical forests are found mainly in the peatlands and lowland of the Amazon, Central Africa, and Southeast Asia, especially in regions of Kalimantan, Sumatra, and Papua New Guinea, where rich biodiversity can still be found and large amounts of carbon are stored in peat soils UNDP, UNEP, WB, and WRI (2000).

Hyperspectral data application for peat forest monitoring in Central Kalimantan, Indonesia

Peatland is a large CO2 reservoir which accumulates 2000Gt of CO2, which is equal to 30% of global soil carbon. However, it has been becoming a large CO2 emission source because of peat decomposition and fire due to drainage water. This is caused by social activities such as canalizing. Especially, in Indonesia, peat swamp forests cover considerable portions of Kalimantan and 37.5% of CO2 emission source is peatland (DNPI, 2010). To take measures, it is necessary to conduct appropriate assessment of CO2 emission in broad peat swamp forest. Because hyperspectral data possess higher spectral resolutions, it is expected to evaluate the detailed forest conditions. We develop a method to assess carbon emission from peat swamp forest by using hyperspectral data in Central Kalimantan, Indonesia. Specifically, we estimate 1) forestry biomass and 2) underground water level expected as an indicator of CO2 emission from peat. In this research, we use the image taken by HyMAP which is one of the airborne hyperspectral sensors. Since the research area differs in forest types and conditions due to the past forest fire and disturbance, forest types are classified with the sparse linear discriminant analysis. Then, we conduct a biomass estimation using Normalized Difference Spectral Index (NDSI). We also analyze the relationship between underground water level and Normalized Difference Water Index (NDWI), and find the possibility of underground water level estimation with hyperspectral data. We plan to establish a highly developed method to apply hyperspectral sensor to peatland monitoring system.

ASTER target observation scenario

12 The Advanced Spaceborne Thermal Emission and Reflectance Radiometer (ASTER) is a multispectral imaging radiometer with 14 spectral bands of VNIR, SWIR and TIR, 60 km imaging swath, and 15-90 m spatial resolution. It was launched on NASAs Terra (EOS AM-1) on December 18, 1999. The ASTER scheduling algorithm and the ASTER Scheduler software was developed in order to maximize the scientific content of each schedule. The Scheduler divides each day into a series of short timesteps (several seconds), for the purpose of prioritization. Prioritization is the process of ranking possible observations, so that the observation opportunities with higher scientific or programmatic value are given higher probabilities of being scheduled. The Scheduler uses the prioritization function to calculate a priority for each potential observation. The prioritization function uses information from all data acquisition request parameters requesting a possible observation. After calculating all the priority, the Scheduler generates 24 hour schedule, namely One Day Schedule (ODS). At each point in this process, the Scheduler checks to make sure that no operating constraints are being violated. Finally, the ODS is transmitted to EOS Operation Center (EOC) every day.