Feasibility Investigation on Determining Soluble Solids Content of Peaches Using Dielectric Spectra
- PDF / 939,981 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 29 Downloads / 171 Views
Feasibility Investigation on Determining Soluble Solids Content of Peaches Using Dielectric Spectra Xinhua Zhu 1 & Lijie Fang 1 & Jingsi Gu 1 & Wenchuan Guo 1
Received: 11 August 2015 / Accepted: 28 October 2015 # Springer Science+Business Media New York 2015
Abstract To investigate the potential of dielectric spectra in determining soluble solids content of intact peaches during postharvest, dielectric constants and dielectric loss factors of 200 intact ‘Hongmi’ peaches were measured at 101 discrete frequencies from 20 to 4,500 MHz using a vector network analyzer and an open-ended coaxial-line probe. Based on the joint x–y distance sample set partitioning (SPXY) method, 160 apples were selected for the calibration set, and the other 40 samples were used for the prediction set. Least squares support vector machine (LSSVM), extreme learning machine (ELM), and back propagation neural network (BPNN) modeling methods were used to establish nonlinear models to predict soluble solids content (SSC) of peaches. To simplify models, 60 and 4 characteristic variables were selected by uninformative variable elimination method (UVE) based on partial least squares and successive projection algorithm (SPA), respectively, and the full dielectric spectra were compressed to seven principal components by principal component analysis (PCA). ELM combined with PCA had the best SSC calibration and prediction performances with predicated correlation coefficient of 0.6986 and predicted root-meansquare error of 0.7763. The poor determination performance indicates that it is difficult to precisely determine soluble solids content of peaches using dielectric spectra.
* Xinhua Zhu [email protected]; [email protected] * Wenchuan Guo [email protected] 1
College of Mechanical and Electronic Engineering, Northwest A&F University, Yangling, Shaanxi 712100, China
Keywords Peach . Dielectric constant . Dielectric loss factor . Soluble solids content . Least squares support vector machine . Extreme learning machine . Back propagation neural network
Introduction Peach (Prunus persica) is the third temperate tree fruit species behind apple and pear in the world (Byrne et al. 2012). It originated in China and has been cultivated for more than 3, 000 years (Capitani et al. 2013). At present, China is the largest producer of peach. Its peach cultivation area and production in 2010 reached 719.4 km2 and 1.04 million tons, respectively, which accounts for about 46 % of the world production (Byrne et al. 2012; Lü et al. 2012). Over 90 % of peach production is for the fresh market. Internal qualities, especially sweetness, determine the fresh fruit price, shelf life, and purchasing desire of consumers. The sweetness is usually determined by soluble solids content (SSC). Traditionally, SSC is measured with an Abbe refractometer or digital refractometer on juice extracted from the fruit. The major shortcoming of this method is its destructive nature. Rapid and nondestructive sensing techniques that can be adapted to on-line processes, such as sorting
Data Loading...
