DEVELOPMENT OF POWER LINE DETECTION ALGORITHM USING FRANGI FILTER AND FIRST-ORDER DERIVATIVE OF GAUSSIAN

Authors

  • P.O Oyibo Department of Computer Engineering, Ahmadu Bello University, Zaria
  • O. Ajayi Department of Computer Engineering, Ahmadu Bello University, Zaria
  • I. Abdulwahab Department of Electrical Engineering, Ahmadu Bello University, Zaria
  • M.T Ogedengbe Department of Mathematics, Statistic/ Computer Science, University of Agriculture, Makurdi
  • A.S Abubakar Department of Electrical Engineering, Ahmadu Bello University, Zaria

Keywords:

Frangi Filter and First-order Derivative of Gaussian (FF-FDOG), Matched Filter-FDOG (MFFDOG), Power Line Detection, True Positive Rate and False Positive Rate.

Abstract

This paper presents the development of a Power Line detection (PLD) algorithm using
Frangi filter and first-order derivative of Gaussian (FF-FDOG). Vision based power line
detection is highly important for threat avoidance in low-altitude flight and also in the
surveillance and maintenance of electrical infrastructure. However, the need for high and
real time detection rates as well as low false alarm in noisy and cluttered images makes it
a challenging task. In this paper the FF-FDOG based threshold was developed using
frangi filter (which detects vessel based on the eigenvalue analysis of the second order
structure of an image) and FDOG filter. Images from the University of South Florida
Computer Vision and Pattern Recognition Group wire database were used to evaluate the
performance of the developed FF-FDOG method. The result obtained was compared to
that obtained when matched filter-FDOG (MF-FDOG) was used for the power line
detection. From the results obtained it was observed that the developed FF-FDOG method
was more efficient with higher True Positive Rate (TPR) (86.39%), and lower False
Positive Rate (FPR) (11.45%) compared to MF-FDOG’s 84.16% TPR and 17.91% FPR.

Published

2019-06-27

Issue

Vol. 8 No. 1 (2019)

Section

Articles