Abstract
In this paper, a multiple classification method for plant species based on Gabor filters and local binary patterns (LBP) operator is proposed. We classify plant species by extracting global texture features with Gabor filters and local features with LBP operator. Simply speaking, the LBP operator is conducted on the magnitude maps of multi-scale and multi-orientation Gabor filtered image rather than original image. Thus, a plant leaf is presented as more spatial histograms with varying scales and orientations. The method has impressively improved the performance comparing with using Gabor transform or LBP absolutely.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have