Identification of Oxtongue Flowers using Image Processing Technique

Authors

  • Yousef Abbaspour-Gilandeh University of Mohaghegh Ardabili
  • Sajad Sabzi 1 Department of Biosystems Engineering, Faculty of Agricultural Technology and Natural Resources, University of Mohaghegh Ardabili, Ardabil, Iran 2Researcher of Research and Innovation Center of ETKA Organization, Tehran, Iran.
  • Farnaz Hoseini Young Researchers and Elite club, Ardabil Branch, Islamic Azad University, Ardabil, Iran.

Abstract

The Oxtongue plant has abundant nutritional properties, and it is consumed worldwide. Identification of flowers from other parts of the plant using image processing is the main step in manufacturing automatic Oxtongue harvesting machine. Our objective is to find suitable color spaces for flower identification by color features. In this study, images were captured in three different conditions, under direct sunshine, in the shadow and under black background. The distance between crop and camera for these three states was 25 cm. Light intensity for under direct sunshine, in the shadow and under black background states were 1800, 407 and 407 lux respectively. After imaging, four color spaces (RGB, HSV, YCbCr and NTSC) were used to determine the suitable color space to identify flowers. For condition under direct sunshine, in the shadow and under black background, HSV, YCbCr and YCbCr were recognized suitable color spaces, respectively. Threshold and run time algorithm for these spaces were H<0.5, S<0.3, V<0.45 and 8.999s, Cb<150 and 7.284s and Cb>150 and 7.204s, respectively. The identification accuracy rate for imaging under direct sunshine, in the shadow and under black background were calculated 100%, 94% and 97.37%, respectively.

Downloads

Published

2015-10-08

Issue

Section

Articles