Method for nondestructively extracting rice spike character through X-ray-visible light dual-mode imaging

Abstract

The invention provides a method for nondestructively extracting rice spike character through X-ray-visible light dual-mode imaging. The method comprises the following steps: synchronously acquiring arice spike reflected light surface image and a transmission light image through a dual-mode imaging system, extracting rice spike neck, spike grain contour, rice grain contour, spike branch skeleton and like characteristics by applying ellipse detection, watershed segmentation, refining processing and Hough transform, obtaining a fusion image through image registration and establishing a mathematical characterization of the rice spike yield trait; representing the total spike grain number by adopting mark numbers of the rice spike grain visible light segmentation image communicating region; representing the grouting degree of the spike grain by adopting the ratio of a registered rice grain area to the rice grain contour area, thereby distinguishing filled grains and blighted grains, performing effective spike judgment and computing a seed percentage; calibrating through a standard sample, and establishing a thousand seed weight mathematical model by combining the spike grain region area integral based on the linear regression analysis and the least square method; computing the spike length, primary branch number, primary branch average length and secondary branch number through a rice spike skeleton path, and representing the grain setting density by using the ratio of the grain number and the spike length. The method provided by the invention has the advantages that the rick spike yield trait can be quickly and accurately acquired under the condition of not threshing and separating the filled grains and blighted grains, and a feasible path is provided for the application of the dual-mode imaging in the rice spike yield trait nondestructive analysis.

Description

technical field [0001] The invention belongs to machine vision detection technology, and specifically relates to a method for non-destructively extracting rice panicle traits by X-ray-visible light dual-mode imaging, that is, simultaneously obtaining rice panicle reflected light surface images and transmitted light panicle neck nodes and The internal image information of panicles and grains, and the two images are registered and fused, combined with the graphic analysis algorithm to obtain the mathematical representation of rice panicle yield traits, including effective panicle, number of grains per panicle, seed setting rate, thousand-grain weight, panicle length, number of primary branches, primary branch The average length of stems, the number of secondary branches, and the density of grains. Background technique [0002] As one of the world's three major food crops, rice is the staple food for more than half of the world's population [1] , my country is the most importa...

AI Technical Summary

Problems solved by technology

Researchers have done a lot of research on the non-destructive testing of rice ears, but there are no reports on the key yield traits of rice ear seed setting rate and thousand-grain weight, and it is difficult to obtain information on the degree of filling inside rice ears with a single imaging method. , it is impossible to effectively distinguish solid grains and shriveled grains on rice spikes

Images