Plant species identification by using scale invariant feature transform (SIFT) and grid based colour moment (GBCM)

• 發表在: 2013 IEEE Conference on Open Systems, ICOS 2013
• 主要作者: 2-s2.0-84897696295
• 格式: Conference paper
• 語言: English
• 出版: IEEE Computer Society 2013
• 在線閱讀: https://www.scopus.com/inward/record.uri?eid=2-s2.0-84897696295&doi=10.1109%2fICOS.2013.6735079&partnerID=40&md5=311cd49b631663dfb57552daa327048c
• DOI: 10.1109/ICOS.2013.6735079

Plant identification using plant leaves is a very challenging task. The most important and crucial phase in plant identification is the phase of feature extraction. This paper presents a method of shape feature extraction that is Scale Invariant Feature Transform (SIFT) and colour feature extraction Grid Based Colour Moment (GBCM) to identify plant. Forty plant species images were collected from their natural habitats and captured under various time of the day. These plant images are then used as ground truth images. These images are further rotated and scaled to produce another forty test images. The extracted features of the test images are then identified by calculating their Euclidean Distance (ED) against the ground truth and achieved identification accuracy rate of 87.5 percent. The proposed feature extraction methods showed potential in identifying plant images captured under natural illumination. However, further work need to be done to improve accuracy of plant identification. © 2013 IEEE.