Machine vision technique for rapid measurement of soybean seed vigor

Authors: LEE, HOONSOO - Us Forest Service (FS); HUY, TRAN QUOC - Chungnam National University; PARK, EUNSOO - Chungnam National University; BAE, HYUNG-JIN - Chungnam National University; BAEK, INSUCK - University Of Maryland; Kim, Moon; MO, CHANGYEUN - Korean Rural Development Administration; CHO, BYOUNG0KWAN - Chungnam National University

Submitted to: Journal of Biosystems Engineering
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/1/2017
Publication Date: 9/15/2017
Citation: Lee, H., Huy, T., Park, E., Bae, H., Baek, I., Kim, M.S., Mo, C., Cho, B. 2017. Machine vision technique for rapid vigor measurement of soybean seed. Journal of Biosystems Engineering. https://doi.org/10.5307/JBE.2017.42.3.227.
DOI: https://doi.org/10.5307/JBE.2017.42.3.227

Interpretive Summary: Seed vigor is an important measure of seed quality, and is determined by both germination rate and seedling growth rate. Vigor testing predicts the general ability of a seed lot to germinate normally over a range of adverse conditions and uses specifications based on speed and uniformity of seedling growth. Because vigor and germination tests are laborious and time-consuming procedures, the development of alternative methods using new technologies is highly sought. This study developed a low-cost computer vision system and morphological image analysis algorithm for acquiring images of seventeen germinating soybean seed samples and their roots during development in a 25°C and 90% relative humidity chamber over a seven-day period. Each day, image processing algorithms were used to extract measures of the following four morphological parameters that are relevant to seed vigor: primary root length, total root length, root area, and diameter (area / total length) from the images. Analysis results showed that the morphological parameters were closely correlated to seed vigor, and that the computer vision algorithm for morphological image analysis has great potential for fast and cost-effective seed sorting and phenotyping for seed vigor prediction to benefit industrial seed producers and processors.

Technical Abstract: Purpose: Morphological properties of roots are important indicators of the vigor of soybean seed, which determine the survival rate of initial growth. Current vigor test of soybean seed relies on manual measurement with human vision. This study describes an application of machine vision technique for rapid vigor measurement of soybean seed to overcome the time-consuming and labor-intensive conventional method. Methods: A CCD camera was used to obtain color image of seed during germination. Image processing techniques such as color spacing, binarization, noise reduction, dilation, skeleton method were used to obtain root segmentation. The various morphological parameters, such as primary root length, total root length, total surface area, average diameter, branching points of roots were calculated from a root skeleton image using a customized pixel-based image processing algorithm. Results: The measurement accuracy of the machine vision system ranged from 92.6% to 98.8% with the accuracy of 96.2% for primary root length and 96.4% for total root length compare to the manual verification. The correlation coefficient was 0.999 with the standard error of prediction of 1.16 for primary root length and 0.97 mm for total root length. Conclusions: The developed machine vision system shows good performance for the morphological measurement of soybean roots. It is assumed that the image analysis algorithm combined with the simple color camera could be used as an alternative method for the conventional seed vigor test method.