Skip to main content
ARS Home » Pacific West Area » Pullman, Washington » Grain Legume Genetics Physiology Research » Research » Publications at this Location » Publication #312247
Title: Image-Based Rapid Phenotyping Method of Chickpeas Seed Size Characterization

Author
item SANKARAN, SINDUJA - Washington State University
item WANG, MENG - Washington State University
item Vandemark, George

Submitted to: Engineering in Agriculture, Environment and Food
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/15/2015
Publication Date: 1/5/2016
Citation: Sankaran, S., Wang, M., Vandemark, G.J. 2016. Image-Based rapid phenotyping method of chickpeas seed size characterization. Engineering in Agriculture, Environment and Food. 9:50-55.

Interpretive Summary: Chickpea was one of the first crops to be domesticated by Neolithic cultures in the Near East 12,000-8,000 years ago. Chickpeas are typically grown in rotation with small cereal grains such as wheat and barley, which benefits cereal production by disrupting cereal disease cycles and making available to cereal crops residual nitrogen produced by beneficial bacteria that colonize chickpea roots. The value of a chickpea crop is influenced by both total seed yield and also by the size of the harvested seed. Larger seeds are used for canning, salads, and fresh markets and have a higher value than smaller seeds, which are typically processed into hummus. Over a recent three year period (2011-2013), large chickpeas have received an average price of $0.39 per pound, which is 51% higher than the average price of $0.25 per pound received for small chickpeas. Seed size in chickpea is typically determined by separating seeds using sieves of different sizes. Both manual and mechanical sieving is laborious and time-consuming, and can sometimes damage the seeds. Our objective was to develop a rapid method to determine the size of chickpea seeds. This method consists of a normal digital camera and a computer that uses a program developed to determine the size of individual chickpea seeds. The program can also determine the average seed size within a sample and the distribution of different seed sizes among a sample. We compared seed size results based on sieving with results obtained using the new imagining method for over 140 samples of chickpea seeds collected from two different field locations. The two methods produced very similar results, which suggest that the new imaging method is at least as accurate as sieving. The image processing technique is a rapid and precise method for determining chickpea size that offers advantages over sieving techniques, which require weighing the samples that stay on top of each screen and often rely on manual shaking, which can be inconsistent across samples. In addition, the image based technique does not result in the destruction of seeds as may occur with shaking and sieving. Chickpea processors may benefit from adopting the image-based technique to determine the distribution of seed sizes within a commercial sample, and this method can likely be applied to other grain legumes including peas and lentils.

Technical Abstract: The value of a chickpea crop is influenced by both total seed yield and also by the size of the harvested seed. Larger seeds are used for canning, salads, and fresh markets and have a higher value than smaller seeds, which are typically processed into hummus. The standard method for determining seed size distribution in grains; the sieve analysis, is labor intensive and time consuming, and can sometimes damage the seeds. Our objective was to develop a rapid method to determine the size of chickpea seeds. An image-based method was developed in this study and was assessed with a large dataset. Samples from a total of 144 plots from two different locations were harvested and seed size was analyzed. Seed size computed from this method was compared to ground-truth data based on the conventional sieving method. The results show that seed size calculated from image-based method was highly correlated to the ground-truth data, with an average correlation coefficient of 0.90. The image processing technique provides a rapid and non-destructive method for determining chickpea seed size that may be of benefit to seed processors. This method can likely be applied to other grain legumes including peas and lentils.