Classification of Bamboo Species by Fourier and Legendre Moment

Bamboo has being widely used as building material throughout the world. From traditional buildings to innovative architectural projects, bamboo has shown its suitability based on a combined low weight, high strength, beauty and durability. The properties of these species vary significantly. A successful application of bamboo in engineering firstly relies on the selection of a correct species. Therefore recognition of bamboo species is necessary before its efficient utilization. Species level identification of bamboos is a highly technical job done primarily by a systematic botanist based on morphological characteristics. However, recognition of the same can also be performed by computer. The bamboo Culm sheath shapes provide valuable data in identification of bamboo species. Automated recognition of bamboo has not yet been well established mainly due to lack of research in this area, non-availability and difficulty in obtaining the database. Using digital image processing and pattern recognition techniques, a supervised classification procedure of three different bamboo species has been developed. In the proposed work, an automated bamboo species recognition system based on shape features of bamboo Culm sheath has been developed using Fourier and Legendre moment classifier. A confusion matrix is created to quantify the class wise and the classifier accuracy. The performance of the classifier is compared based on the classifier accuracy and classwise accuracy. It is concluded the Fourier moment have significantly good results than the Legendre moment. The system can eliminate the need for laborious human recognition method requiring a plant taxonomist. The results obtained shows considerable recognition accuracy proving that the techniques used is suitable to be implemented for commercial purposes.