Submitted to: Proceedings of SPIE
Publication Type: Proceedings
Publication Acceptance Date: October 30, 2009
Publication Date: March 12, 2010
Citation: Kandala, C., Sundaram, J., Hinson, J.W. 2010. Estimation of kernels mass ratio to total in-shell peanuts using low-cost RF impedance meter. Proceedings of SPIE. Interpretive Summary: It would be very useful if we can estimate, at least to a certain extent, the amount of mature peanut kernels we may obtain when a bag of peanuts are shelled prior to their shelling. This would be particularly useful for the buyer. An attempt was made here, to use low cost Chari Impedance meter measurements for this purpose. The conduction of RF signals through a medium depends on the density of the medium and the variation of certain electrical properties of a parallel-plate capacitor holding a dielectric medium (peanuts in this case) could be used to estimate the relative compactness of the medium. The ratio of the weight of mature kernels to the weight of the in-shell peanuts (weight ratio) was correlated to the capacitance, impedance and phase angle values and a calibration equation was developed. From this equation, weight ratio of any unknown sample could be determined nondestructively. By this method the percentage weight ratio could be determined within 1% of their reference values for over 90% of the samples tested.
Technical Abstract: In this study estimation of percentage of total kernel mass within a given mass of in-shell peanuts was determined nondestructively using a low-cost RF impedance meter. Peanut samples were divided into two groups one the calibration and the other the validation group. Each group contained 25 samples of about 100 g of peanuts. Capacitance, phase angle and impedance measurements on in-shell peanut samples were made at frequencies 1 MHz, 5 MHz and 9 MHz. Ten measurements on each sample set were made, to minimize the errors due to the orientation of the peanuts as they settle between the electrodes of the impedance meter, by emptying and refilling the samples after each measurement. After completing the measurements on each set the peanuts from that set were shelled, kernels were separated and weighed. Multi linear regression (MLR) calibration equation was developed by correlating the percentage of the kernel mass in a given peanut sample set with the measured capacitance, impedance and phase angle values. This equation was used to predict the kernel mass ratio of the samples from the validation group. The fitness of the MLR equation was verified using Standard Error of Prediction (SEP) and Root Mean Square Error of Prediction (RMSEP). Also the predictability of total kernel mass ratio was calculated by comparing the mass ratio predicted using MLR model with the actual mass ratio determined using the conventional standard method of visual determination.