Determination of the Vibration Response of Sugarcane Stalk to Predict Fiber and Brix Using Image Processing

Sugarcane is one of the most important products in the area of agriculture. The solid part of sugarcane, fiber (bagasse), is the major byproduct of sugarcane industry and one of the largest cellulose sources. Another measurable index in sugarcane is Brix which is considered as sugar content as well as maturity index. Due to the importance of fiber, Brix index and moisture content of sugarcanes’ stalk, in this paper has been tried to provide nondestructive method for estimation of aforementioned parameters. To this end mechanical vibration of sugarcane stalks was evaluated as a cantilever beam. After placing samples in suitable place, they were excited and natural frequency, damping and damping ratio was calculated; these features were given to artificial neural network as inputs to predict brix, fiber percent and moisture content of sugarcane’s stalk (as outputs). The correlation coefficient for fiber, brix and moisture content were R2=0.97, R2=0.71 and R2=0.55 respectively. The support vector machinewas used to classify sugarcane parameters. The results showed that support vector machine classify fiber and Brix with the accuracy of 91.43 and 83.73%, respectively.