Browsing by Author "Ahmadu, Hussein Adomu"

Now showing 1 - 4 of 4
  • Item
    EVALUATING THE PERFORMANCE OF EPIPOLAR GEOMETRY FOR AUTOMATIC IMAGE REGISTRATION
    (Department of Geography, University of Ilorin, 2017-04) Ahmadu, Hussein Adomu; Abdulyekeen, Abdulaziz Ola; Alade, Azeez Kola; Yusuf, Abdulganiyu
    Image registration has being of great importance to a variety of applications in remote sensing, medical diagnosis, computer vision, and pattern recognition. Accordingly, several methods have been proposed by different researchers, with a view to enhance the performance of image registration process. This paper investigates the performance of automatic image registration using epipolar geometry. The epipolar geometry was recovered through the estimation of a fundamental matrix from matched image points. Then the rectification transform was computed using the F-matrix, and the images registered using epipolar correlation. A root mean square error (RMSE) of 0.032 and 00 was recorded for the correlated points. the paper reveals that the image registration process performed well and therefore suitable for remote sensing image registration
  • Item
    EVALUATING THE POTENTIAL OF NIKON D3200 AND CANON EOS 7D AMATEUR CAMERAS FOR 3D RECONSTRUCTION
    (Nigerian Institution of Surveyors, 2021-10) Ahmadu, Hussein Adomu; Idrees, Mohammed Oludare; Ayanwale, Victor Opeyemi; Omar, Dahir Muazu; Ayo, Babalola
    The growing use of amateur cameras for several photogrammetric applications such as point determination, three-dimensional reconstruction, measurements and archeological documentation, have ignited studies on the accuracy potential of these categories of cameras. This paper investigates the accuracy potential of Nikon D3200 and Canon EOS 7D amateur cameras for 3D reconstruction of the oldest hostel building in University of Ilorin, Nigeria. Before image collection, a total number of 37 points, 9 control points for image processing and 28 check points for accuracy assessment were strategically marked and coordinated on the building using Sokkia set 520 total Station. Thereafter, multiple overlapping images of the building facade were taken successively with the two cameras in such a way that adjacent pair have end lap of between 80 - 85%. In Agisoft Photoscan, after removing bad images, the photos were aligned. Subsequently, the 3D mesh and textured models wee generated. Corresponding model points of the 28 measured check points were extracted for accuracy determination using the root mean square error (RMSE) analysis. The results showed that both cameras have acceptable point accuracy with the Nikon D3200 producing RMSE of 2.55mm, 3.23mm, 2.51mm and Canon EOS 7D yielded RMSE of 2.77mm, 4.76mm, 2.68mm, in xyz respectively.
  • Item
    Finding lasting solutions to the early deterioration of constructed roads in Nigeria.
    (Nigerian Journal of Technology, 2019-10) Ahmadu, Hussein Adomu; Abdulyekeen, Abdulazeez Ola; Alade, Azeez Kola; Yusuf, Abdulganiyu
    The socioeconomic benefits of having good roads is hardly realized in most developing countries, due to the fact that most constructed roads tend to deteriorate very rapidly before their expected life span. Several factors have been identified by different researchers to be the cause of this problem. This paper tried to find solutions to some of these factors using three phases of route survey and design of roadways, namely; route location, data acquisition, and geometric design. The results showed that most of the factors that facilitate early deterioration of constructed roads could be solved through the processes involved in the three phases of route survey and design of roadways. The paper concluded that route survey and design of roadways thoroughly addresses and solves most of the factors that facilitate early deterioration of constructed roads.
  • Item
    An Investigation into the performance and accuracy of Automatic Image Registration Using Epipolar Geometry
    (Department of Geography and Environmental Management, University of Ilorin., 2018-07-03) Ahmadu, Hussein Adomu; Abdulyekeen, Abdulazeez Ola; Alade, Azeez Kola; Yusuf, Abdulganiyu
    Image registration has being of great importance to a variety of applications in remote sensing, medical diagnosis, computer vision and pattern recognition. Owing to this, several methods have been adopted by different researchers, with a view to enhance the performance and accuracy of image registration process. This paper presents an investigation into the performance of automatic image registration using epipolar geometry. The epipolar geometry was recovered through the estimation of a fundamental matrix from matched image points, rectification transforms were computed using the F-matrix, and the images were registered using epipolar correlation. A Root mean square error (RMSE) of 0.032 and 0.00 was recorded for the correlated points, and the paper concluded that the performance of the image registration process was successful and accurate.