The main problem posed by Polarimetric Synthetic Aperture Radar (PolSAR) image classification in remote sensing is the ability to develop classifiers that can substantially discern the different classes inherent in natural and man-made targets. Emphasis has shifted from the use of conventional classifiers to modern non-parametric classifiers such as the Artificial Neural Network (ANN) and Support Vector Machine (SVM); and most recently the hybrid Deep Neural Network (DNN) which is a fusion of Deep Learning (DL) and ANN. This research therefore presents the novel application of Deep Support Vector Machine (DSVM) which is a fusion of DL and SVM to PolSAR image classification. Two PolSAR images of Flevoland region in Netherlands, and Winnipeg in Canada are used as test beds for the experiment. The Lee filter is used to filter the images to suppress the speckle noise in the images. The Pauli decomposition is applied to decompose the images into |𝑆𝑆 𝐻𝐻𝐻𝐻 + 𝑆𝑆 𝑉𝑉𝑉𝑉 |, |𝑆𝑆 𝐻𝐻𝐻𝐻 − 𝑆𝑆 𝑉𝑉𝑉𝑉 |, |𝑆𝑆 𝐻𝐻𝑉𝑉 | polarimetric channels. Then the Gray Level Co-occurrence Matrix (GLCM) texture feature for |𝑆𝑆 𝐻𝐻𝐻𝐻 + 𝑆𝑆 𝑉𝑉𝑉𝑉 |, |𝑆𝑆 𝐻𝐻𝐻𝐻 − 𝑆𝑆 𝑉𝑉𝑉𝑉 |, |𝑆𝑆 𝐻𝐻𝑉𝑉 | are extracted based on correlation, contrast, energy, and homogeneity statistics, using GLCM directions 0 0 , 45 0 , 90 0 , and 135 0 with an offset distance of 60. To enhance the efficiency of the model 8, 16, 32, 64, 128, and 256 quantization levels are explored. The DSVM classifier is implemented with four kernel function: Exponential Radial Basis Function (ERBF), Gaussian Radial Basis Function (GRBF), neural, and polynomial. The first set of results is a comparison of the DSVM and SVM.

No abstract

Efficient transformation parameters are key to effective geodetic operations between any systems. With the advancement in technology, an improved interaction between geodetic reference frames has been noticed irrespective of their inherent heterogeneous deformations and the intermittently present oversight that exist as a result of the conversion methodology.The spatial data captured using the Global Navigation Satellite System (GNSS) has a reference datum based on the World Geodetic System 1984 (WGS84) ellipsoid. These data usually require a transformation to a local projection with its ellipsoid and datum and vice versa. These geodetic operations are unavoidably vulnerable to data loss due to distortion, especially if the applied model is unsuitable for the transformation between reference frames. Therefore, the main aim of this study is to authenticate the existing set of parameters for the local and geocentric systems in Nigeria by assessing the efficacy of the Bursa Wolf (BW) and the Molodensky Badekas (MB) models. To this end, both models (BW and MB) are compared in this study and used in the development of the Helmerts seven transformation parameters between Minna datum (Clarke 1880) and WGS84 reference ellipsoids for a large region in Southern Nigeria. Results show that the sets of datum shift transformation parameters for both models (a scale factor with three sets of translational and rotational parameters) derived from the exercise of the BW and the MB models revealed a high degree of correlation between the expected and derived set of coordinates during the validation/testing phase using each set of models. The validation exercise was carried out on a total of seven points, which were not part of the original computations. These points were distributed across the region to provide a better framework anda higher confidence level. Overall, an improvement of 68% is observed in terms of the correlation between expected and derived coordinates in the validation pointsand from the obtained root mean square values. Ultimately, the MB model is preferred to the BW model evidently because the transformation involves a global-local reference frame. .

No abstract

Lagos has undergone an unprecedented urban expansion. Contemporary findings favour the integration of cellular automata and geographic information systems for modelling land use change. This research introduces the support vector machine based GIS cellular automata calibration for land use change prediction of Lagos. The support vector machine based cellular automata model is loosely coupled with the geographic information systems. Support vector machine parameters are optimised with the k-fold cross-validation technique, using the linear, polynomial, and RBF kernels functions. The land use change prediction is based on three land use epochs: 1963-1978, 1978-1984, and 1984-2000. The performance of the model was evaluated using the Kappa statistic and receiver operating characteristic. The order of performance of the three kernels is: RBF, polynomial, and linear. The results indicate substantial agreement between the actual and predicted maps. The urban forms in 2015 and 2030 are predicted based on the three land use epochs.