No abstract

Wildland fire is one of the most causes of deforestation, and it has an important impact on atmospheric emissions, notably CO2. It occurs almost every year in Indonesia, especially during the dry season. Therefore, it is necessary to identify the burned areas from remote sensing images to establish the zoning map of areas prone to wildland fires. Many methods have been developed for mapping burned areas from low-resolution to medium-resolution satellite images. One of the popular approaches for mapping tasks is a deep learning approach using U-Net architecture. However, it needs a large amount of representative training data to develop the model. In this paper, we present a new dataset of burned areas in Indonesia for training or evaluating the U-Net model. We delineate burned areas manually by visual interpretation on Landsat-8 satellite images. The dataset is collected from some regions in Indonesia, and it consists of 227 images with a size of 512 × 512 pixels. It contains one or more burned scars or only the background and its labeled masks. The dataset can be used to train and evaluate the deep learning model for image detection, segmentation, and classification tasks related to burned area mapping.

In a modern survey, information on the real condition of the study area is required to support the analysis and interpretation result of a study. However, obtaining information on the real condition in a wide covered area is difficult, particular in an area that hard to access and has varied topographic. The method that can imaging the real condition of a study area is observation using UAV/drone using structure from motion technique. Besides can be observed with a wide area, the detailed condition of the area also can be visualized. Structure from motion (sfm) is the technique that determines the spatial and geometric relationship of the target area through the movement of the camera. In this research, the sfm technique was applied to create the 3 dimension construction of the Kelok Sembilan flyover. The result show, 3D construction has a high spatial resolution in 2.99 cm/pixel measured in Ground Sampling Distance (GSD). Meanwhile, the horizontal relative resolution is 5.97 cm, and the vertical relative resolution is 8.95 cm.

Speckle is noise found in SAR data that will affect the image interpretation process. To reduce the presence of speckle in SAR data, a speckle filtering process is needed. This study will evaluate the multi temporal speckle filtering for Sentinel-1 image, VH polarisation. The data used in this study are 30 SENTINEL-1 images recorded in different seasons in 2018. The methods applied for multi temporal speckle filters are Boxcar, Frost, Lee, Gamma Map and Lee Sigma with window sizes of 5x5 and 7x7. The quality test for speckle filter results will be done qualitatively (by looking at the visual appearance of the filter results) and quantitatively (Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and Normalized Mean (NM). The best qualitative and quantitative multitemporal speckle filter results are obtained from the Frost Filter 5x5 with a minimum of 3 scene data.

High resolution images data from Terrasar-X are used to extract digital elevation model (DEM) using stereo radargrammetry in the attempt to achieve better resolution of terrain surface in Indonesia. As sample in this study, stereo pairs images from TerraSAR-X StripMap mode (~3m resolution) on Madiun city is used with difference of incidence angle around ~18.88 degree to extract the elevation of the area. Furthermore, field observation on the selected area will be used on elevation accuracy assessment. The digital surface elevation (DSM) generated by stereo radargrammetry in this study shows us high resolution with spatial pixel spacing 5.57 meter and elevation accuracy around ~4 meter.