OpenUCT :: Browsing by Author "Sithole, George"

Browsing by Author "Sithole, George"

Now showing 1 - 11 of 11

Open Access
Analysing the road reserve encroachment in Maseru Lesotho using remote sensing and image analysis
(2021) Ralitsoele, Teboho; Sithole, George
The increasing rate of urbanization and the problem of road reserve encroachment mean that there is no space for road expansion and sometimes for maintenance and road furniture, these and other problems have exposed the problem of road reserve encroachment. The main aim of this study was to investigate methods of finding the road reserve encroachment in Maseru Lesotho using aerial photos. The study used single image analysis and multiple image analysis methods. In single image analysis, the study used three methods of image classifications to find objects that are in the road reserve. Under classification, the study used both supervised and unsupervised image classifications. For supervised classification, the study used the direct image classification method where the aim was to look for every object found in the road reserve. For the indirect approach, the study looked for the ground to find objects in the road reserve. For unsupervised image classification, the study assumed that small clusters are encroachment. In multiple images analysis, the study used the 2015 and 2017 images to determine permanent objects found to have encroached road reserves. Here the assumption was that encroachment does not change over time, which means that unchanged objects during the change detection have encroached on the road reserve. The confusion matrix was used to tell the best performing method and the results show that the indirect method, both in Qoaling and Maqalika performed best. All the methods showed that there was an encroachment on a road reserve, and found that permanent objects were; houses, shops, and shopping centers. The study recommended the use of images with higher resolution and more bands, also that images be taken frequently.
Open Access
Assessing the accuracy of openstreetmap data in south africa for the purpose of integrating it with authoritative data
(2014) Siebritz, Lindy-Anne; Sithole, George
The introduction and success of Volunteered Geographic Information (VGI) has gained the interest of National Mapping Agencies (NMAs) worldwide. VGI is geographic information that is freely generated by non-experts and shared using VGI initiatives available on the Internet. The NMA of South Africa i.e. the Chief Directorate: National Geo- Spatial Information (CD: NGI) is looking to this volunteer information to maintain their topographical database; however, the main concern is the quality of the data. The purpose of this work is to assess whether it is feasible to use VGI to update the CD: NGI topographical database. The data from OpenStreetMap (OSM), which is one the most successful VGI initiatives, was compared to a reference data set provided by the CD: NGI. Corresponding features between the two data sets were compared in order to assess the various quality aspects. The investigation was split into quantitative and qualitative assessments. The aim of the quantitative assessments was to determine the internal quality of the OSM data. The internal quality elements included the positional accuracy, geometric accuracy, semantic accuracy and the completeness. The _rst part of the qualitative assessment was concerned with the currency of OSM data between 2006 and 2012. The second part of the assessment was focused on the uniformity of OSM data acquisition across South Africa. The quantitative results showed that both road and building features do not meet the CD: NGI positional accuracy standards. In some areas the positional accuracy of roads are close to the required accuracy. The buildings generally compare well in shape to the CD: NGI buildings. However, there were very few OSM polygon features to assess, thus the results are limited to a small sample. The semantic accuracy of roads was low. Volunteers do not generally classify roads correctly. Instead, many volunteers prefer to class roads generically. The last part of the quantitative results, the completeness, revealed that commercial areas reach high completeness percentages and sometimes exceed the total length of the CD: NGI roads. In residential areas, the percentages are lower and in low urban density areas, the lowest. Nonetheless, the OSM repository has seen signi_cant growth since 2006. The qualitative results showed that because the OSM repository has continued to grow since 2006, the level of currency has increased. In South Africa, the most contributions were made between 2010 and 2012. The OSM data set is thus current after 2012. The amount and type of contributions are however not uniform across the country for various reasons. The number of point contributions was low. Thus, the relationship between the type of contribution and the settlement type could not be made with certainty. Because the OSM data does not meet the CD: NGI spatial accuracy requirements, the two data sets cannot be integrated at the database level. Instead, two options are proposed. The CD: NGI could use the OSM data for detecting changes to the landscape only. The other recommendation is to transform and verify the OSM data. Only those features with a high positional accuracy would then be ingested. The CD: NGI currently has a shortage of sta_ that is quali_ed to process ancillary data. Both of the options proposed thus require automated techniques because it is time consuming to perform these tasks manually.
Open Access
Context based detection of urban land use zones
(2011) Louw, Johan; Sithole, George
This dissertation proposes an automated land-use zoning system based on the context of an urban scene. Automated zoning is an important step toward improving object extraction in an urban scene.
Open Access
Elements of design for indoor visualisation
(2016) Gangraker, Muhammad Zaid; Sithole, George
Indoor visualisation has received little attention. Research related to indoor environments have primarily focussed on the data structuring, localisation and navigation components (Zlatanova et al., 2013). Visualisation is an integral component in addressing the diverse array of indoor environments. In simple words, 'What is the most efficient way to visualise the surrounding indoor environment so that the user can concisely understand their surroundings as well as facilitating the process of navigation?' This dissertation proposes a holistic approach that consists of two components. The significance of this approach is that it provides a robust and adaptable method in providing a standard to which indoor visualisation can be referenced against. The first component is a theoretical framework focussing on indoor visualisation and it comprises of principles from several disciplines such as geovisualisation, human-perception theory, spatial cognition, dynamic and 3D environments as well as accommodating emotional processes resulting from human-computer interaction. The second component is based on the theoretical framework and adopts a practical approach towards indoor visualisation. It consists of a set of design properties that can be used for the design of effective indoor visualisations. The framework is referred to as the "Elements of Design" framework. Both these components aim to provide a set of principles and guidelines that can be used as best practices for the design of indoor visualisations. In order to practically demonstrate the holistic indoor visualisation approach, multiple indoor visualisation renderings were developed. The visualisation renderings were represented in a three-dimensional virtual environment from a first-person perspective. Each rendering used the design framework differently. Also, each rendering was graded using a parallel chart that compares how the different visual elements were used per the rendering. The main findings were that the techniques/ renderings that used the visual elements effectively (enhanced human-perception) resulted in better acquisition and construction of knowledge about the surrounding indoor environment.
Open Access
An empirical assessment of real-time progressive stereo reconstruction
(2016) Westaway, Matthew; Sithole, George
3D reconstruction from images, the problem of reconstructing depth from images, is one of the most well-studied problems within computer vision. In part because it is academically interesting, but also because of the significant growth in the use of 3D models. This growth can be attributed to the development of augmented reality, 3D printing and indoor mapping. Progressive stereo reconstruction is the sequential application of stereo reconstructions to reconstruct a scene. To achieve a reliable progressive stereo reconstruction a combination of best practice algorithms needs to be used. The purpose of this research is to determine the combinat ion of best practice algorithms that lead to the most accurate and efficient progressive stereo reconstruction i.e the best practice combination. In order to obtain a similarity reconstruction the in t rinsic parameters of the camera need to be known. If they are not known they are determined by capturing ten images of a checkerboard with a known calibration pattern from different angles and using the moving plane algori thm. Thereafter in order to perform a near real-time reconstruction frames are acquired and reconstructed simultaneously. For the first pair of frames keypoints are detected and matched using a best practice keypoint detection and matching algorithm. The motion of the camera between the frames is then determined by decomposing the essential matrix which is determined from the fundamental matrix, which is determined using a best practice ego-motion estimation algorithm. Finally the keypoints are reconstructed using a best practice reconstruction algorithm. For sequential frames each frame is paired with t he previous frame and keypoints are therefore only detected in the sequential frame. They are detected , matched and reconstructed in the same fashion as the first pair of frames, however to ensure that the reconstructed points are in the same scale as the points reconstructed from the first pair of frames the motion of the camera between t he frames is estimated from 3D-2D correspondences using a best practice algorithm. If the purpose of progressive reconstruction is for visualization the best practice combination algorithm for keypoint detection was found to be Speeded Up Robust Features (SURF) as it results in more reconstructed points than Scale-Invariant Feature Transform (SIFT). SIFT is however more computationally efficient and thus better suited if the number of reconstructed points does not matter, for example if the purpose of progressive reconstruction is for camera tracking. For all purposes the best practice combination algorithm for matching was found to be optical flow as it is the most efficient and for ego-motion estimation the best practice combination algorithm was found to be the 5-point algorithm as it is robust to points located on planes. This research is significant as the effects of the key steps of progressive reconstruction and the choices made at each step on the accuracy and efficiency of the reconstruction as a whole have never been studied. As a result progressive stereo reconstruction can now be performed in near real-time on a mobile device without compromising the accuracy of reconstruction.
Open Access
Enhancing point cloud processing using audio cues
(2014) Ntsoko, Thabo; Sithole, George
Today many airborne and terrestrial acquisitions capture point clouds of scenes or objects to be modelled. But before modelling can be done point clouds need to be taken through processing steps such as registration, cleaning, simpli_cation, etc. These point clouds are usually manually processed before being processed automatically. Manual processing of point clouds depends on the visual interaction the user has with the point cloud provided by the visual cues. This research investigated enhancing the level of interaction the user has with the point cloud when processing it. The proposed method augments audio in point clouds to enhance its processing where visual cues are limited. This investigated _nding objects/points of interest in the point cloud while processing it by estimating the position (azimuth and elevation) and depth of audio objects associated with these point cloud objects. The occupancy of space of audio objects was also investigated to determine the unseen events around objects of interest in the point cloud. For example, in a scan registration problem, audio could be augmented to a misaligned scan. As this scan is manually rotated and translated into alignment, various audio cues can be used to inform the user of the state of this alignment. An outlier separated from a surface in a point cloud could be identi_ed and removed by augmenting audio to a volumetric brush that does the point cloud cleaning. Associating audio cues of the audio object with the depth of the outlier to the surface could help the user identify this outlier. Similar implementation could be adopted in point cloud simpli_cation tasks. Various audio cues exist which allow a listener to discern particular information about a sound source. This is done by the human auditory system, using cues such as intensity, pitch, reverberation and HRTFs to discern this information. However, limitations exist in retrieving this information. Literature supports the use of the auditory interface in applications commonly built for the visual interface. The addition of the auditory interface is seen as a way of increasing the interaction users have with applications and therefore improving the experience. An auditory interface was built to help undertake this research. The test subject was immersed in the auditory environment by wearing headphones. This meant that the subject and the virtual listener were merged, allowing the subject to receive emitted audio. The perception of the audio was with respect to the virtual listener.
Open Access
Establishing the kinematics of the North East Region of South Africa with the use of GPS data
(2015) Desai, Ahmad; Sithole, George; Combrinck, Ludwig
The Nubia-Somalia boundary is one of the least well defined plate boundaries. The southern extension, as it extends through the latitudes that are co-incidental to South Africa's geographical extent has not been defined. This project analyses Trignet data, supplemented with other available Nubian GPS data, to determine whether the North East Region of South Africa is segregated from the Nubian Plate. A ten year GPS time series (processed using the Bernese 5.0 precise GNSS software) is minimally constrained to the ITRF2008, using the reprocessed IGS08 products. This processed data was managed to alleviate GPS time series bias, which includes seasonal effects, the determination of GPS velocity uncertainty using a white+ icker noise model, discontinuity detection using Roggero's (2012) discontinuity algorithm and an outlier detection using an algorithm developed for this project for each of the 104 stations incorporated in the network. The computed velocities were compared to the ITRF2008 published velocities and a RMS of the post t residuals of 0.4, 0.4 and 0.7 mm/yr for the North, East and Height were computed respectively. This RMS is of similar magnitude to the WRMS of the post t velocity residuals in computing transformation parameters from ITRF2005 to ITRF2008, suggesting that the threshold of ITRF2008 velocity precision has been obtained. The resultant Euler vector was scrutinised on three fronts: an Euler vector rep- resenting the geographical extent of South Africa, an Euler vector in which the height velocity was minimised to zero and an Euler vector representing the North East Region of South Africa. All computed Euler vectors are consistent, suggesting a uniform South African GPS velocity field. The Euler vector representative of South Africa's uniform velocity field has an angular velocity of 0.273 deg/Myr, with a pole of rotation at lat 50.85 and long -82.83 degrees and is consistent with the Euler vectors computed by Altamimi et al. (2012) and Malservisi et al. (2013) for Africa and South Africa respectively. The consistency of the Euler vector suggests that South Africa is underlain by a single plate (i.e. the Nubian Plate), and that the East African Rift valley does not extend into South Africa. Hence, for future tectonic studies and long term reference frame solutions, stations such as HARB, HRAO and RBAY should be included, as they are not located in deformation zones.
Open Access
Multiple scenario analyses forecasting the impacts of sea level rise in Cape Town, South Africa
(2018) Taukoor, Sheveenah Sunnassee; Ansorge, Isabel; Sithole, George
Sea level rise is highly interdisciplinary and its study entails not only oceanography, but other fields such as geomatics, climatology and geology. In this study we relied on the tools from geomatics to produce sea level rise maps in order to assess the vulnerability of the coastline of Cape Town, South Africa. After generating a DEM of a spatial resolution of 2 m from LiDAR point cloud data, we made use of GIS to design 4 sea level rise scenarios based on the RCP 2.6, RCP 4.5, RCP 6.0 and RCP 8.5 scenarios from the IPCC. Among the findings, it was found that 2.16 – 3.09 km² of land would be potentially inundated by 2100. The main receptors which were identified were sandy beaches, rocky shores and built-up land. Permanent inundation would possibly change the appeal and the nature of the beaches and affect the tourism industry. Hence the coastline requires immediate attention as it is one of the most valuable assets in the tourism industry. Tidal effect and storm surge effect were also identified as additional factors which brought temporary changes to the sea level in Cape Town. These impacts were further investigated in 8 coastal suburbs (Tableview, Woodbridge Island, Paarden Eiland, Foreshore, Sea Point, Glencairn, Fish Hoek and Strand.) Suitable adaptation strategies including hard protection measures (e.g groynes, sea walls, barriers) and soft protection measures (e.g beach nourishment) were also proposed for these 8 suburbs.
Open Access
Reconstruction of industrial piping installations from laser point clouds using profiling techniques
(2009) Mapurisa, Willard T; Sithole, George
As-built models of industrial piping installations are essential for planning applications in industry. Laser scanning has emerged as the preferred data acquisition method of as built information for creating these three dimensional (3D) models. The product of the scanning process is a cloud of points representing scanned surfaces. From this point cloud, 3D models of the surfaces are reconstructed. Most surfaces are of piping elements e.g. straight pipes, t-junctions, elbows, spheres. The automatic detection of these piping elements in point clouds has the greatest impact on the reconstructed model. Various algorithms have been proposed for detecting piping elements in point clouds. However, most algorithms detect cylinders (straight pipes) and planes which make up a small percentage of piping elements found in industrial installations. In addition, these algorithms do not allow for deformation detection in pipes. Therefore, the work in this research is aimed at the detection of piping elements (straight pipes, elbows, t-junctions and flange) in point clouds including deformation detection.
Open Access
Urban scene description for a multi scale classication of high resolution imagery case of Cape Town urban Scene
(2012) Ikokou, Guy Blanchard; Sithole, George
In this paper, a multi level contextual classification approach of the City of Cape Town, South Africa is presented. The methodology developed to identify the different objects using the multi level contextual technique comprised three important phases.
Open Access
Vine Signal Extraction - an Application of remote sensing in precision Viticulture
(2010) Smit, Julian; Sithole, George; Strever, A E
This paper presents a study of precision agriculture in the wine industry. While precision viticulture mostly aims to maximise yields by delivering the right inputs to appropriate places on a farm in the correct doses and at the right time, the objective of this study was rather to assess vine biomass differences. The solution proposed in this paper uses aerial imagery as the primary source of data for vine analysis. The first objective to be achieved by the solution is to automatically identify vineyards blocks, vine rows, and individual vines within rows. This is made possible through a series of enhancements and hierarchical segmentations of the aerial images. The second objective is to determine the correlation of image data with the biophysical data (yield and pruning mass) of each vine. A multispectral aerial image is used to compute vegetation indices, which serve as indicators of biophysical measures. The results of the automatic detection are compared against a test field, to verify both vine location and vegetation index correlation with relevant vine parameters. The advantage of this technique is that it functions in environments where active cover crop growth between vines is evident and where variable vine canopy conditions are present within a vineyard block.

Browsing by Author "Sithole, George"

Results Per Page

Sort Options