Device and method for detecting stored-grain insects based on visible light-near infrared binocular machine vision

Abstract

The invention relates to a device and a method for detecting stored-grain insects, which consists of a stored-grain insect separation part, a stored-grain insect transmission part and a machine vision part, wherein the stored-grain insect separation part efficiently separates stored-grain insects from a grain sample by utilizing a screening motor to drive a vibrating screen; the stored-grain insect transmission part can transmit an acquisition box to an appointed position and can automatically clear screen underflows in the acquisition box; the machine vision part consists of a control module, a computer, a lighting box, a near infrared camera, a visible light CCD, a light source and corresponding software. The method for detecting stored-grain insects comprises the following steps of sequentially acquiring near infrared images and visible light images of the screen underflows for information fusion, extracting entire morphological characteristics and local morphological characteristics of the stored-grain insects, and automatically determining the kinds and the quantity of the live stored-grain insects by utilizing an optimized stored-grain insects characteristic space and identification software, thus realizing the real-time, precise and automatic detection of normal stored-grain insects.

Description

technical field [0001] The invention relates to a detection device and method for storage pests, in particular to an automatic detection method and device for storage pests based on visible light-near infrared binocular computer vision. Background technique [0002] my country is the largest grain production, storage and consumption country in the world, so doing a good job in grain storage is undoubtedly more important. The grain weight loss after harvesting in the world is about 10-15%, and at least 5% of the grain during the annual storage is spoiled by pests. The loss rate of my country's treasury grain reserves is about 0.2%, which is a remarkable achievement. However, the price paid for this is also huge, especially the extensive use of pesticides. In order to ensure the safe storage of grain, stored grain must be fumigated with pesticides at least once a year, and in many places, fumigated twice or even more, which not only increases expenses, but also pollutes food...

AI Technical Summary

Problems solved by technology

Qiu Daoyin, Zhang Hongtao and others have developed an online intelligent detection system for grain insects for 9 types of grain insects. Currently, the system can count and automatically classify grain insects in real time, but live insects will be repeatedly counted. In addition, grain samples are directly sampled without screening which affects the efficiency of the system

[0009] The machine vision detection method can realize the automatic classification of grain insects, which is the development direction of grain insect detection. However, the current machine vision system can classify and identify up to 9 types of grain insects, and cannot distinguish the "dead" and "alive" grain insects. , unable to overcome the "fake death" problem, and the types of recognition need to be further increased

Images