Thesis (M. Tech.) -- Central University of Technology, Free State, 2009 / Automated Guided Vehicles (AGVs) have many different design specifications, although they all have certain design features in common, for instance they are designed to follow predetermined paths, and they need to be aware of their surroundings and changes to their surroundings. They are designed to house sensors for navigation and obstacle avoidance. In this study an AGV platform was developed by modifying an electric wheelchair. A serial port interface was developed between a computer and the control unit of the electric wheelchair, which enables the computer to control the movements of the platform. Different sensors were investigated to determine which would be best suited and most effective to avoid collisions. The sensors chosen were mounted on the AGV and a programme was developed to enable the sensors to assist in avoiding obstacles. An imaging device as an additional sensor system for the AGV was investigated. The image produced by a camera and dome mirror was processed into a panoramic image representing an entire 360o view of the AGV‟s surroundings. The reason for this part of the research was to enable the user to make corrections to the AGV‟s path if it became stuck along the track it was following. The entire system was also made completely wireless to improve the flexibility of the AGV‟s applications.

Geospatial analysis involves application of statistical methods, algorithms and information retrieval techniques to geospatial data. It incorporates time into spatial databases and facilitates investigation of land cover (LC) dynamics through data, model, and analytics. LC dynamics induced by human and natural processes play a major role in global as well as regional scale patterns, which in turn influence weather and climate. Hence, understanding LC dynamics at the local / regional as well as at global levels is essential to evolve appropriate management strategies to mitigate the impacts of LC changes. This can be captured through the multi-resolution remote sensing (RS) data. However, with the advancements in sensor technologies, suitable algorithms and techniques are required for optimal integration of information from multi-resolution sensors which are cost effective while overcoming the possible data and methodological constraints. In this work, several per-pixel traditional and advanced classification techniques have been evaluated with the multi-resolution data along with the role of ancillary geographical data on the performance of classifiers. Techniques for linear and non-linear un-mixing, endmember variability and determination of spatial distribution of class components within a pixel have been applied and validated on multi-resolution data. Endmember estimation method is proposed and its performance is compared with manual, semi-automatic and fully automatic methods of endmember extraction. A novel technique - Hybrid Bayesian Classifier is developed for per pixel classification where the class prior probabilities are determined by un-mixing a low spatial-high spectral resolution multi-spectral data while posterior probabilities are determined from the training data obtained from ground, that are assigned to every pixel in a high spatial-low spectral resolution multi-spectral data in Bayesian classification. These techniques have been validated with multi-resolution data for various landscapes with varying altitudes. As a case study, spatial metrics and cellular automata based models applied for rapidly urbanising landscape with moderate altitude has been carried out.

Indiana University-Purdue University Indianapolis (IUPUI) / The primary objective of this research involves mapping out and validating the existence of sub surface drainage tiles in a given cropland using Remote Sensing and GIS methodologies. The process is dependent on soil edge differentiation found in lighter versus darker IR reflectance values from tiled vs. untiled soils patches. Data is collected from various sources and a primary classifier is created using secondary field variables such as soil type, topography and land Use and land cover (LULC). The classifier mask reduces computational time and allows application of various filtering algorithms for detection of edges. The filtered image allows an efficient feature recognition platform allowing the tile drains to be better identified. User defined methods and natural vision based methodologies are also developed or adopted as novel techniques for edge detection. The generated results are validated with field data sets which were established using Ground Penetration Radar (GPR) studies. Overlay efficiency is calculated for each methodology along with omission and commission errors. This comparison yields adaptable and efficient edge detection techniques which can be used for similar areas allowing further development of the tile detection process.