Method for detecting and grading internal quality of fruits

Abstract

The invention relates to a method for detecting and grading internal quality of fruits based on a support vector machine (SVM) model and near infrared spectrum (NIRS) detection technology. The method is characterized by introducing the concept of degree of membership to quality analysis of the fruits aiming at the ambiguity of the fruit taste, simultaneously considering the NIRS characteristics and physically and chemically measured quality parameters of the fruits in the process of determining the degree of membership of fruit quality determining and adopting the degree of membership based on support vector domain description to correct the degree of membership of determining which is determined based on the sugar degree and the acidity, thereby establishing the effective SVM model for fruit grading. The non-destructive detecting and grading method based on the SVM model and the NIRS detection technology not only realizes grading of the fruit sample quality by the traditional methods but also can measure the degree of membership of the samples to the subordinate quality grade via the degree of membership and can obtain higher grading precision according to the difference between the predicted value and the measured value of the degree of membership of the fruit samples, thereby better recognizing the internal quality of the fruits.

Description

technical field [0001] The invention relates to a method for detecting and grading the internal quality of fruits, in particular to a method for detecting and grading the internal quality of fruits based on near-infrared spectrum detection technology and a support vector machine model (ie SVM model). Background technique [0002] Fruit non-destructive testing refers to the testing of internal quality parameters such as sugar content, acidity, and hardness of fruit without destroying the shape of the fruit. These indicators have an important impact on the grading and pricing of fruits. Fruits have certain spectral characteristics, and detecting and analyzing their spectral characteristics (that is, the selection, absorption and reflection properties of fruits to light) is an important aspect of fruit non-destructive testing. [0003] The interaction process between light and fruit is: when a beam of light hits the fruit, the fruit will reflect, transmit and absorb the light. ...

AI Technical Summary

Problems solved by technology

This ambiguity is reflected in the lack of clear judgment standards, that is, it is impossible to use uniform and established standards to clearly classify the taste of fruits

At the same time, this ambiguity also makes it difficult to obtain deterministic evaluation results when using traditional grading methods to classify the taste of fruits.

Images