Multi remote sensing data in landslide detection and modelling

Landslide is one of the disasters that threaten the human’s lives and properties in mountainous environments like Malaysia with high elevation and steep terrain. Mitigation and prediction of this phenomenon can be done through the detection of landslide areas. Therefore, an appropriate landslide analysis method is needed in order to map and consequently understand the characteristics of landslide disaster. This reasearch adopted several approaches to investigate, analyze, and assess landsliding in terms of detection, modeling and optimization of the landslides conditioning factors. Remote sensing (RS) and geographic information system (GIS) techniques can support overall landslides management as they can produce rapid data collection and analysis for hazard studies. Therefore, current research is divided into two general aspects. The first aspect which mainly utilized RS technology is to detect the landslides areas using active microwave sensor of ALOS Palsar sensor. Active radar data have been broadly used for hazard and especially landslides mapping due to its precision in detection of landslide areas. Active remote sensing sensors can provide their own illumination source and they can record data independent of day and night time. Another advantage is their capability to penetrate the cloud cover, making the image recording independent of all weather conditions. Gunung pass area, Malaysia was used as case study to detect the landslides using interferometric synthetic aperture RADAR (InSAR) generated from ALOS-PALSAR repeat pass data. The results were validated using the observed reference point of the landslides and the root mean square error (RMSE) was 0.19. Furthermore, advance 3D processing was performed for measuring the volume of the landslides. Additionaly, the ascending orbit ALOS PALSAR images were acquired from September 2008, January 2009 and December 2009 to generate the DInSAR to model the horizontal movement. Subsequently the displacement measurements of the study site (Gunung Pass) were calculated. The accuracy of the result was evaluated through its comparison with ground truth data using the R2 and RMSE methods. The resulted deformation map showed the landslide locations in the study area from interpretation of the results with 0.84 R2 and 0.151 RMSE. DInSAR precision was 11.8 cm which proved the efficiency of proposed method in detecting landslides in tropical country like Malaysia. On the other hand, the data fusion technique are used between LiDAR airborne laser scanner data (high density) and high resolution QuickBird imagery (2.6m spatial resolution) to map the landslide events in Bukit Antarabangsa, Ulu Klang, Malaysia. Wavelet transformer (WT) technique was utilized to perform the fusion. Furthermore,this research employed the Taquchi technique for optimization of the segmentation parameters. Moreover, rule-based technique was performed for object-based classification. Confusion matrix was used to examine the proficiency and reliability of the proposed method. The achieved overall accuracy and kappa coefficient are 90.06% and 0.84 respectively. In addition, the direction of the mass movement was recognized by overlaying the final classification map with LiDAR-derived slope and aspect factors. The second aspect of the current research is related to the GIS spatial modeling. For all proposed landslide susceptibility methods such as EBF and SVM, landslides inventory was provided and randomly divided into two datasets; 70% for training the models and the remaining 30% was used for validation purpose. Subsequently the related conditioning factors’ datasets were constructed and utilized in the analysis. Some researchers assume that as the number of conditioning factors increases, the accuracy of the generated susceptibility map increases. By contrast, other case studies prove that a small number of conditioning factors are sufficient to produce landslide susceptibility maps with a reasonable quality. This study investigates the effects of conditioning factors on landslide susceptibility mapping. Bukit Antarabangsa, Ulu Klang, Malaysia was selected as the study area,because it is a catchment area with a high potential of landslide occurrence. A spatial database of 31 landslide locations was evaluated to map landslide-susceptible areas. Two datasets of conditioning factors were constructed to be used in the analysis. The first dataset was derived from high-resolution airborne laser scanning data (LiDAR),which contains eight landslide conditioning factors such as altitude, slope, aspect,curvature, stream power index (SPI), topographic wetness index (TWI), topographic roughness index (TRI), and sediment transport index (STI). The second dataset was gathered using the same conditioning factors of the first dataset, but with the addition of other conditioning factors: geological and environmental factors of soil, geology,land use/cover (LULC), distance from river, and distance from road. Two different datasets were constructed to compare the efficiency of one over the other in landslide susceptibility zonation. Three different types of methods were implemented to recognize the importance of different conditioning factors in landslide mapping.weights-of-evidence (WoE) (bivariate statistical analysis (BSA)), logistic regression (LR) (multivariate statistical analysis (MSA)), and data-driven support vector machine (SVM) were used to determine the optimal landslide conditioning factors. The area under curve (AUC) was used to assess the obtained results. The prediction rates of WoE, LR, and SVM obtained from only the LiDAR-derived conditioning factors were 59%, 86%, and 84%, respecti vely. The prediction rates of the WoE, LR, and SVMobtained from the second dataset were 65%, 66%, and 69%, respectively