Reduction of Variable Illumination Effect on Pixel Gray-levels of Machine Vision

  • Suh S. R. (Bio-Systems and Agric. Engineering. Dept. Inst. of Agric. Sci & Tech., Chonnam National University) ;
  • Huang J. K. (Bio-Systems and Agric. Engineering. Dept. Inst. of Agric. Sci & Tech., Chonnam National University) ;
  • Kim Y. T. (Bio-Systems and Agric. Engineering. Dept. Inst. of Agric. Sci & Tech., Chonnam National University) ;
  • Yoo S. N. (Bio-Systems and Agric. Engineering. Dept. Inst. of Agric. Sci & Tech., Chonnam National University) ;
  • Choi Y. S. (Bio-Systems and Agric. Engineering. Dept. Inst. of Agric. Sci & Tech., Chonnam National University) ;
  • Sung J. H. (National Institute of Agricultural Engineering, RDA)
  • Published : 2004.06.01

Abstract

This study was carried out to develop methods of reducing the effect of solar illumination on pixel gray-levels of machine vision for agricultural field use. Two kinds of monochrome CCD cameras with manual and auto-iris lenses were used to take pictures within a range of 15 to 120 klux of solar illumination. A camera having more precise automatic control functions gave much better result. Four kinds of indices using pixel gray-level of the $99\%$ white DRS (diffuse reflectance standard) as a reference were tried to compensate pixel gray-levels of an image for variable illumination. Coefficients of variation of the indices within a range of illumination were used as a criterion for comparison. The study concluded that an index of (A+B)/A, where A is gray-level of the $99\%$ DRS and B is gray-level of the tested material, gave the best consistency in the range of solar illumination.

Keywords