Engineering characterization of selected cereal grain and legume crops for physical quality determination using machine vision technology

Sharma, Rohit

Engineering characterization of selected cereal grain and legume crops for physical quality determination using machine vision technology

Date

2021

Authors

Sharma, Rohit

Publisher

Punjab Agricultural University, Ludhiana

Abstract

A study was undertaken for the determining relevant engineering properties for physical quality estimation of selected wheat and green gram refractions using machine vision technology. The most common wheat varieties PBW 725 and Unnat PBW 343 and green gram varieties SML 832 and SML 668 were considered for the present study. The kernels were grouped into four categories such as sound grains, damaged grains, shriveled grains and broken grains for wheat. The type of grains for green gram include sound grains, shriveled grains, discolored grains and broken grains. For image acquisition, Digital images of one hundred kernels of each selected refraction of wheat and green gram were acquired using the flatbed scanner model CanoScan 5600F with CCD 6-line color with white fluorescent light source. The computer programming language Python version 3.7 and OpenCV image processing library were used for analysis of the images. A total of 30 features related to size and shape (14 features), color (10 features) and texture (6 features) of the selected kernels. For each category of refractions, red (R), green (G), and blue (B) color densities were extracted. Algorithms have been developed for analysis of images of refraction. A linear relationship was obtained between actual and predicted lengths and widths of selected refractions. The experimental results for length of refractions were found non-significant in comparison to predicted value at 95% confidence interval with R2 in the range of 0.81–0.94. The comparison results for width for all refractions are found statistically non-significant at 95% confidence interval and R2 was in the range of 0.73–0.92 for wheat refractions of both selected varieties. The paired samples t-test results showed that all parameters determined with image processing method was not significantly (P>0.05) different from the same parameters measured with vernier caliper. Another linear relationship was found between individual kernel weight and projected area with R2 in the range of 0.51 to 0.92 and between the volume of refractions derived from measured dimensions and calculated from image with R2 in the range of 0.75 to 0.93 for both wheat varieties. Correlation behavior between different parameters were observed using pearson correlation matrix. It was observed that in all wheat and green gram refractions, the individual kernel weight was found correlated significantly with projected area, perimeter, major diameter, feret diameter arithmetic mean diameter, geometric mean diameter, surface area and volume of kernels.

Citation

Sharma, Rohit (2021). Engineering characterization of selected cereal grain and legume crops for physical quality determination using machine vision technology (Unpublished Ph.D. Dissertation). Punjab Agricultural University, Ludhiana, Punjab, India.