Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Development of portable nondestructive detection device for mango internal diseases

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  2022 ASABE Annual International Meeting  2200171.(doi:10.13031/aim.202200171)
Authors:   Yang Li, Yankun Peng, Yongyu Li, Decai Lv, Le Liu
Keywords:   Portable device; Mango; Internal diseases; Non-destructive detection; Visible/near infrared spectroscopy

Abstract. Mango internal diseases are one of the main factors affecting mango quality. Using traditional methods to detect internal diseases in mangoes requires cutting open mangoes to determine the disease status, which has problems such as low detection efficiency and destructiveness. Based on visible/near infrared spectroscopy technology, it is of great value to develop a portable mango internal diseases detection device to realize convenient, nondestructive, rapid and accurate detection of mango internal diseases. The device was mainly composed of light sources, a shell, a fruit holder and a data acquisition and analysis module, etc. The transmission spectrum data of 30 mangoes with internal diseases and 30 healthy mangoes were collected, respectively. The samples were randomly divided into calibration set and prediction set in a ratio of 3:1. Based on the original spectrum, standard normal variate transform (SNV), multivariate scattering correction (MSC) and SG-smoothing (SGS) preprocessed spectra, the partial least squares - discriminant analysis models (PLS-DA) of mango disease conditions were established in the range of 450nm-950nm. The results showed that the SGS-PLS-DA method had the best modeling effect, the sensitivity, specificity, and classification accuracy rate of correction set were 100.00%, 100.00% and 100.00%. The sensitivity, specificity, and classification accuracy rate of prediction set were 100.00%, 87.50% and 93.33%. In order to eliminate irrelevant information and improve model prediction performance. Based on the SGS preprocessing, the competitive adaptive reweighed sampling (CARS) algorithm was used to select the characteristic wavelengths of the spectrum to establish PLS-DA prediction model. The modeling results showed that the SGS-CARS-PLS-DA model had the same prediction effect with SGS-PLS-DA, but it was reduced from 1497 wavelength points to 12 wavelength points, reducing the amount of data by 99.2%. Therefore, the portable mango internal diseases detection device and related modeling methods developed based on visible/near infrared spectroscopy technology could easily, quickly, non-destructively and accurately detect mango internal diseases, and provided effective technology support for internal diseases detection in mango planting and sales.

(Download PDF)    (Export to EndNotes)