Nondestructive testing method for content of total sugar in blueberry fruits based on deep learning

Abstract

The invention discloses a nondestructive testing method for the content of total sugar in blueberry fruits based on deep learning. According to the method, blueberry fruits are classified by using deep learning; the accuracy and the efficiency of blueberry fruit product detection are improved; the method comprises the following steps: collecting blueberry fruits in different maturation periods; determining the anthocyanin content and the total sugar content of the blueberries; then establishing a pericarp pigment content prediction network SPCPN based on anthocyanin and blueberry image correlation; establishing a fruit internal quality prediction network FIQPN based on anthocyanin and total sugar correlation; finally, merging the peel pigment content prediction network and the fruit internal quality prediction network into the blueberry quality parameter prediction network BQPPN. After external verification, the prediction of the network on the total sugar content of the blueberry fruits is greater than 94%, and the method has the advantages of no damage, high efficiency, high accuracy, convenience and stable result.

Description

technical field [0001] The invention belongs to the technical field of intelligent detection of fruit sugar content, and in particular relates to a nondestructive detection method for total sugar content in blueberry fruit based on deep learning. Background technique [0002] With the improvement of people's living standards, blueberries have attracted much attention and favor because of their unique flavor and strong nutritional and health functions. Blueberries are native to North America, and China has a short history of cultivation, but the growth rate is very fast, reaching more than 20,000 hectares by 2012. China's blueberries are currently mainly exported. As the proportion of blueberries put on foreign markets increases each year, the requirements for blueberry fruit quality are gradually increasing. However, there are many limitations in product grading only by artificial eyes, and traditional fruit quality testing is mostly used. The chemical experiment method is ...

AI Technical Summary

Problems solved by technology

At present, domestic and foreign research on the non-destructive detection of total sugar content in blueberry fruit mostly focuses on the use of near-infrared light detection technology and has become mature, but there is no case of using deep learning to detect the total sugar content in blueberry fruit. The main reason is It is the traditional computer vision method that has no breakthrough progress before deep learning, and the accuracy of using computer vision methods in fruit quality inspection is very limited

Images