No abstract

Abstract:In this study, a method based on supervised machine learning is proposed to identify village buildings from open high-resolution remote sensing images. We select Google Earth (GE) RGB images to perform the classification in order to examine its suitability for village mapping, and investigate the feasibility of using machine learning methods to provide automatic classification in such fields. By analyzing the characteristics of GE images, we design different features on the basis of two kinds of supervised machine learning methods for classification: adaptive boosting (AdaBoost) and convolutional neural networks (CNN). To recognize village buildings via their color and texture information, the RGB color features and a large number of Haar-like features in a local window are utilized in the AdaBoost method; with multilayer trained networks based on gradient descent algorithms and back propagation, CNN perform the identification by mining deeper information from buildings and their neighborhood. Experimental results from the testing area at Savannakhet province in Laos show that our proposed AdaBoost method achieves an overall accuracy of 96.22% and the CNN method is also competitive with an overall accuracy of 96.30%.

Abstract. Coastal forests are known to protect coastal areas from environmental degradation. In this paper, we examined another important role of coastal forests -to mitigate tsunami devastations to coastal areas. Using a twodimensional numerical model (Harada and Imamura model, 2005), we evaluated the damping effects of a coastal forest to resist tsunami inundation in Yogyakarta, Indonesia. In the simulations, we set up a two-km long control forest with a representative topography of the study site and experimented its damping performance sensitivity under various width configurations, e.g. 20, 40, 60, 80, 100 and 200 m. The initial tsunami wave was set such that the inundation depth at the front edge of the forest would not exceed 4 m (tree fragility limit). The forest variables such as species, density, DBH, height and canopy size were determined from a typical forest of the site (Casuarina plantation, 4 trees/100 m 2 , Diameter at Breast Height = 0.20 m). The results showed that coastal forest with 100 m width reduced inundation flux, depth and area by 17.6, 7.0 and 5.7 %, respectively. Exponential models were found to describe the relationships between forest width and tsunami inundation transmission. An additional experiment was performed using actual topography and a forest plantation plan with 100 m width for 2.46 km 2 . In this experiment, the results showed that the plan would reduce inundation flux by 10.1 %, while the exponential model estimated it to be 10.6 %, close to the numerical model results. It suggests that statistical models of forest width and damping effects are useful tools for plantation planning, as it allows for quicker evaluation of the impact of coastal forest without simulation modeling that requires a lot of data, time and computing power.

No abstract