Zainuddin Yusoff

Automatic landslide detection using Dempster–Shafer theory from LiDAR-derived data and orthophotos

ABSTRACT A good landslide inventory map is a prerequisite for landslide hazard and risk analysis. In tropical countries, such as Malaysia, preparation of the landslide inventory is a challenging task because of the rapid growth of vegetation. Thus, it is crucial to use rapid and accurate technique and effective parameters. For this purpose, Dempster Shafer theory (DST) was applied in fusing high resolution LiDAR derived data products and Greenness index derived from orthophoto imagery. Two sites were selected, for the implementation and evaluation of the DST model; site “A” for DST implementation and site “B” for the comparison. For model implementation, vegetation index, slope and height were used as effective parameters for identifying automatic landslide detection. Two type of DST based fusions were evaluated; (greenness and height) and (greenness and slope). Furthermore, validation techniques were used to validate the accuracy are confusion matrix and area under the curve. The overall accuracy of the first and second evaluated fusions were (73.4% and 84.33%), and area under the curve were (0.76 and 0.81) respectively. Additionally, the result was compared with Random Forest (RF) based detection approach. The results showed that DST does not require a priori knowledge.

Assessment of Lake Sediments Properties and Contaminations Level

Sediments are principal carriers of the trace elements in the hydrosphere. Properties of the sediment (such as grain size, specific surface area and pore volume) decide the concentration level of the pollutant contain in water. The properties of sediment differed in each lake according to the normal geological phenomenon and source of discharge wastewater. The purpose of this study is to investigate the properties and contamination level of the sediment collected from lakes in Universiti Putra Malaysia (UPM). Sediment samples are taken from three different lakes; Lake IT, Lake ENG and Lake PK. These locations are selected due to the type of wastewater has been discharged into the lakes which are from colleges and academic buildings. The sediments were tested in terms of physical, chemical properties and contaminant concentration (Pb, P and Cu). Using the contaminant concentration results; the sediment concentration level of the pollutant of each lakes were referred to the Consensus- Based Sediment Quality Guidelines (CBSQG-2003). The dominant grain sizes of the sediments were found in the range of silt/clay; with the size fraction in the range 12.74% to 12.83%. The specific surface areas of sediments were in the range of 16.3 to 22.5 m2/g with a pore size distribution in the range of 20 to 29 mm3/g. The chemical properties show that the pH values are in normal range; pH 7, TOC values in the range of 10.84 to 12.39% and salinity values in the range of 0.05 to 0.06 dS/m. The contaminant concentrations show that the main heavy metal presents in Lake IT, Lake ENG and Lake PK as Lead (Pb) with 0.033 mg/l, 0.036 mg/L and 0.038 mg/L, respectively. According to the CBSQG-2003, due to the concentration of Lead presents in lakes sediment in UPM area, the sediments were categorised as non-polluted.

A hybrid model using machine learning methods and GIS for potential rockfall source identification from airborne laser scanning data

The main objectives of this paper are to design and evaluate a hybrid approach based on Gaussian mixture model (GMM) and random forest (RF) for detecting rockfall source areas using airborne laser scanning data. The former model was used to calculate automatically slope angle thresholds for different type of landslides such as shallow, translational, rotational, rotational-translational, complex, debris flow, and rockfalls. After calculating the slope angle thresholds, a homogenous morphometric land use area (HMLA) was constructed to improve the performance of the model computations and reduce the sensitivity of the model to the variations in different conditioning factors. After that, the support vector machine (SVM) was applied in addition to backward elimination (BE) to select and rank the conditioning factors considering the type of landslides. Then, different machine learning methods [artificial neural network (ANN), logistic regression (LR), and random forest (RF) were trained with the selected best factors and previously prepared inventory datasets. The best fit method (RF) was then used to generate the probability maps and then the source areas were detected by combining the slope raster (reclassified according to the thresholds found by the GMM model) and the probability maps. The accuracy assessment shows that the proposed hybrid model could detect the potential rockfalls with an accuracy of 0.92 based on training data and 0.96 on validation data. Overall, the proposed model is an efficient model for identifying rockfall source areas in the presence of other types of landslides with an accepted generalization performance.

Granites in Malaysia: From Hard Rock to Clay Minerals

Tropical areas with extreme climates are host to extreme weathering processes and the weathered materials are normally left in situ with the absence of large-scale denudation processes such as glaciations. This research tries to understand the behaviour of the weathered granites in Malaysia, from hard rock to the final products, the clay minerals. Grade 1 or fresh granites were sampled from different locations in Malaysia and analysed. The residual soil above the fresh granites, which were formed from the weathering activities were also analysed. The types of clay minerals and clay-sized particle grains found from two study locations were compared. The bases of the comparisons were index properties, strength properties and the mineralogical properties. The parent rocks were also analysed to obtain the origin of the minerals formed at the later stages of weathering. It was found that the strength of the soil mass formed from the weathering processes generally depend on the clay-sized particle grains rather than the types of clay minerals. It should however be noted that only halloysites and smectites clay minerals were observed in the samples obtained from the two study locations.

Optimized Hierarchical Rule-Based Classification for Differentiating Shallow and Deep-Seated Landslide Using High-Resolution LiDAR Data

Landslide is one of the most devastating natural disasters across the world with serious negative impact on its inhabitants and the environs. Landslide is considered as a type of soil erosion which could be shallow, deep-seated, cut slope, bare soil, and so on. Distinguishing between these types of soil erosions in dense vegetation terrain like Cameron Highlands Malaysia is still a challenging issue. Thus, it is difficult to differentiate between these erosion types using traditional techniques in locations with dense vegetation. Light detection and ranging (LiDAR) can detect variations in terrain and provide detailed topographic information on locations behind dense vegetation. This paper presents a hierarchical rule-based classification to obtain accurate map of landslide types. The performance of the hierarchical rule set classification using LiDAR data, orthophoto, texture, and geometric features for distinguishing between the classes would be evaluated. Fuzzy logic supervised approach (FbSP) was employed to optimize the segmentation parameters such as scale, shape, and compactness. Consequently, a correlation-based feature selection technique was used to select relevant features to develop the rule sets. In addition, in other to differentiate between deep-seated cover under shadow and normal shadow, the band ration was created by dividing the intensity over the green band. The overall accuracy and the kappa coefficient of the hierarchal rule set classification were found to be 90.41 and 0.86%, respectively, for site A. More so, the hierarchal rule sets were evaluated using another site named site B, and the overall accuracy and the kappa coefficient were found to be 87.33 and 0.81%, respectively. Based on these results, it is demonstrated that the proposed methodology is highly effective in improving the classification accuracy. The LiDAR DEM data, visible bands, texture, and geometric features considerably influence the accuracy of differentiating between landslide types such as shallow and deep-seated and soil erosion types like cut slope and bare soil. Therefore, this study revealed that the proposed method is efficient and well-organized for differentiating among landslide and other soil erosion types in tropical forested areas.

SQUEEZING POTENTIAL EVALUATION OF TUNNEL IN TROPICAL AREA

In recent years, there has been an increasing interest in the tunnel construction. This part describes the squeezing behavior of poor rock mass associated with deformability and strength properties. Squeezing phenomena happen in tunnels which are surrounded by weak and moderately strength of rock. Squeezing cause to deformed the tunnels cross section and wastes a lot of human and natural source in all of the word every year. The purpose of current study is to determine methods employed to classifying and quantifying of potential squeezing in tunnel. The results show that some part of case-study tunnel has potential of squeezing. Along with the empirical and semi-empirical approaches is available in order to evaluating of potential of squeezing in tunnel are presented moreover squeezing potential evaluation of Padang Renas tunnel which is located in tropical area (Malaysia) are presented. . The implications of the anticipated ground conditions and squeezing on machine and ground support selection as well as the field observation of the actual conditions are discussed in this paper.