TY - JOUR
T1 - Color index based thresholding method for background and foreground segmentation of plant images
AU - Castillo-Martínez, Miguel
AU - Gallegos-Funes, Francisco J.
AU - Carvajal-Gámez, Blanca E.
AU - Urriolagoitia-Sosa, Guillermo
AU - Rosales-Silva, Alberto J.
N1 - Publisher Copyright:
© 2020 Elsevier B.V.
PY - 2020/11
Y1 - 2020/11
N2 - In this paper, the color index based thresholding method for background and foreground segmentation of plant images is presented. The proposed method is implemented with color index approach, for this purpose two color indexes are modified to provide better information about the green color of the plants. Two fixed threshold methods are proposed for the color indexes to discriminate between foreground (green plant) and background (soil). Three versions of the proposed method are presented, these are applied in plant images with controlled conditions and crop images with real environmental conditions. Experimental results demonstrate that the proposed method outperforms other algorithms used as comparative in plant images obtaining a segmentation error of 6.62 ± 5.85% and a classification ratio of 1.93 ± 0.05. Also, the proposed method provides better segmentation results in comparison with other well-known state-of-art algorithms in different crop images. Finally, the proposed method does not require of complex calculus and their implementations are straightforward on any device.
AB - In this paper, the color index based thresholding method for background and foreground segmentation of plant images is presented. The proposed method is implemented with color index approach, for this purpose two color indexes are modified to provide better information about the green color of the plants. Two fixed threshold methods are proposed for the color indexes to discriminate between foreground (green plant) and background (soil). Three versions of the proposed method are presented, these are applied in plant images with controlled conditions and crop images with real environmental conditions. Experimental results demonstrate that the proposed method outperforms other algorithms used as comparative in plant images obtaining a segmentation error of 6.62 ± 5.85% and a classification ratio of 1.93 ± 0.05. Also, the proposed method provides better segmentation results in comparison with other well-known state-of-art algorithms in different crop images. Finally, the proposed method does not require of complex calculus and their implementations are straightforward on any device.
KW - Color index
KW - Green plants
KW - Segmentation
KW - Threshold method
UR - http://www.scopus.com/inward/record.url?scp=85090906943&partnerID=8YFLogxK
U2 - 10.1016/j.compag.2020.105783
DO - 10.1016/j.compag.2020.105783
M3 - Artículo
AN - SCOPUS:85090906943
SN - 0168-1699
VL - 178
JO - Computers and Electronics in Agriculture
JF - Computers and Electronics in Agriculture
M1 - 105783
ER -