Hyper-spectral reflection image collecting system and corn seed purity nondestructive detection method based on same

Abstract

The invention discloses a hyper-spectral reflection image collecting system and a corn seed purity nondestructive detection method based on the same. The method comprises the steps that an image of each wave band is extracted as a sub-image of a corresponding hyper-spectral reflection image based on hyper-spectral reflection images with the quantity of M of a collected corn seed sample under wave bands with the quantity of M; a corresponding corn seed outline image and shape characteristic parameters with the quantity of 12M under the M wave bands are extracted by utilizing an image segmentation method as characteristic parameters for judging an affiliated category of the corresponding corn seed sample, a preset detection model is inputted to perform the detection, and a detection result on the affiliated purity category of the corresponding corn seed sample is acquired. Due to the adoption of the hyper-spectral reflection image collecting system and the corn seed purity nondestructive detection method based on the same, defects in the prior art that parameters at the wave band with a loud noise are difficult to extract, the accuracy is poor, the reliability is low and the like can be overcome, and the advantages such as simplicity in operation, good noise resistance, good accuracy, high reliability and good real-time performance can be realized.

Description

technical field [0001] The invention relates to hyperspectral reflection image technology and seed purity detection technology, in particular to a hyperspectral reflection image acquisition system and a non-destructive detection method for corn seed purity based on the system. Background technique [0002] The country is based on agriculture, and agriculture is based on seeds. Seeds are the most basic means of production in agricultural production and are everything. In recent years, due to the continuous advancement of corn hybridization technology, the variety of corn seeds in my country has continued to increase, resulting in a decline in the purity of each variety to a certain extent. With the continuous improvement of people's living standards, consumers will have large differences in the taste of different varieties of corn when purchasing corn. Therefore, improving the purity of corn seeds is the fundamental way to solve this problem. [0003] In response to this pr...

AI Technical Summary

Problems solved by technology

[0004] At present, the maize seed detection technology based on machine vision has been widely used, but the traditional visible light machine vision technology is based on visible light source. The more serious it is, the worse the distinguishability of features will be, thus affecting the accuracy of seed detection

[0005] The hyperspectral reflectance image contains both the visible light region and the near-infrared region. In the hyperspectral image, the gray value displayed by the image in each band is different, and the effective feature information of the corn seeds contained is compared with the traditional machine vision. However, in the noisy band, the traditional histogram threshold segmentation method cannot effectively segment the corn seed contour, and can only segment a few of the bands. Therefore, the hyperspectral reflection image data volume cannot be fully utilized Features of large size and large number of bands

[0006] In the process of realizing the present invention, the inventors found that there are at least some defects in the prior art, such as difficulty in extracting parameters of bands with high noise, poor accuracy and low reliability.

Images