Rice test device based on a convolutional neural network and a use method thereof

Abstract

The invention discloses a rice seed testing device based on a convolutional neural network and a use method thereof, and the use method comprises the following steps: separating a single seed from a rice seed photo, and extracting a seed contour by using a computer vision algorithm; classifying the separated kernels according to the outline pictures of the kernels, and dividing the kernels into closed kernels, closed kernels, open kernels, open kernels, closed kernels, closed kernels, closed kernels, open kernels and open kernels; training the deep convolutional neural network by using variousgrain contour photograph samples; verifying the deep convolutional neural network by using a verification photo; and inputting a to-be-processed rice seed photo into the deep convolutional neural network for test. Compared with the prior art, the method has the advantages that the rice is effectively and quickly divided into eight types, the eight types of rice are counted at the same time, and the problem that seed classification and statistics are not detailed in the test process is solved.

Description

technical field [0001] The invention relates to the field of rice seed testing, in particular to a rice seed testing device based on a deep learning convolutional neural network and a method for using the same. Background technique [0002] In the prior art, generally use machine equipment or a combination of machine equipment and machine vision to test seeds, such as CN201010234207 (digital rice test plant) discloses a digital test plant, which separates solid grains and empty grains by air separation Kernels were classified into intact and damaged kernels using the Bayesian classification method in Machine Vision Technology Dynamics. But there are the following deficiencies: one is that in the process of seed testing, the classification and statistics of seeds are not detailed, such as only being divided into two kinds of solid grains and grains, and there is no classification of seeds such as solid grains, semi-solid grains, batch grains, and half grains. Statistics; the...

AI Technical Summary

Problems solved by technology

[0004] The object of the present invention is to provide a rice seed test device based on convolutional neural network and its use method, to solve the problem of incomplete seed classification and statistics in the process of seed test proposed in the background technology

Images