Using Spectral Decomposition to Detect Dirty Solar Panels and Minimize Impact on Energy Production


  • Ernesto Zamora Ramos University of Nevada, Las Vegas
  • Suzanna Ho University of Nevada, Las Vegas
  • Evangelos A Yfantis University of Nevada, Las Vegas



Solar Panel, Photovoltaic Cell, Pattern Recognition, Mahalanobis Distance, Classifier


Dirt and dust deposits on the surface of a solar panel array obstruct the amount of light that can reach the photovoltaic cells, reducing the amount of electricity produced. Solar panels are cleaned when the energy drop has already occurred and is detected. This work presents an algorithm designed to detect dirty solar panels. It is based on the spectral decomposition of the scattered light reflected off the panels' surface by analyzing color images of the surface obtained using digital cameras. It applies the statistical classification method of Mahalanobis distance to separate images where it detects the excess reflected light, classifying them as having a high probability of representing dirty solar panels. It aims to minimize the loss of energy by warning solar plants operators to clean panels before the energy drop becomes significant.

Author Biographies

Ernesto Zamora Ramos, University of Nevada, Las Vegas

Computer Science Department

Suzanna Ho, University of Nevada, Las Vegas

 Computer Science Department

Evangelos A Yfantis, University of Nevada, Las Vegas

Computer Science Department


(1) K. W. Böer, Solar Cells. (2015) Chemistry Explained [Online]. Available:

(2) A. J. McEvoy, T. Markvart and L. Castañer, Principles of Solar Cell Operation. Practical handbook of photovoltaics: fundamentals and applications, 2nd ed., Academic Press, Elsevier, 2012, pp. 7–30.

(3) C. P. Ryan, F. Vignola, and D. K. McDaniels, Solar cell arrays: Degradation due to dirt. Proceedings of 1989 Annual Conference of The American Solar Energy Society, 1989, pp. 234-237.

(4) M. K. Mazumder et al., Solar Panel Obscuration by Dust and its Mitigation in the Martian Atmosphere. Particles on Surfaces 9: Detection, Adhesion and Removal, 2006, pp. 1-29.

(5) G. J. McLachlan, Mahalanobis Distance. Resonance, Jun. 1999, pp. 20-26.

(6) E. Zamora Ramos, Using Image Processing Techniques to Estimate the Air Quality. McNair Scholars Research Journal, UNLV chapter, 6th ed., 2012, pp 189-194.

(7) E. W. Weisstein, Central Limit Theorem. [Online]. Available:

CentralLimitTheorem.html. [Accessed 15 Nov 2015].

(8) R. Maitra, Discrimination and Classification – Introduction. [Online]. Available:, 2012. [Accessed 30 May 2014].

(9) E. A. Yfantis et al., Pollution Detection in Urban Areas Using the Existing Camera Networks. International Journal of Multimedia Technology, 2013, Vol. 3, No. 3, pp. 98-102.

(10) E. A. Yfantis and A. Fayed, A Camera System for Detecting Dust and Other Deposits on Solar Panels. Journal Of Advances in Image and Video Processing, 2014, Vol. 2, No. 5, pp. 1-10.




How to Cite

Zamora Ramos, E., Ho, S., & Yfantis, E. A. (2016). Using Spectral Decomposition to Detect Dirty Solar Panels and Minimize Impact on Energy Production. European Journal of Applied Sciences, 3(6), 01.