Fissure Detection and Measurement in Rough Rice Using X-Ray Imaging

Abstract. Fissures in rice kernels that develop prior to harvest and post-harvest processing significantly reduce head rice yield, a crucial parameter for evaluating rice quality and economic value in the rice industry. In this study, fissures in rough rice were revealed by scanning approximately 50 rough rice kernels at a time using an x-ray system. An algorithm was developed to detect and measure fissures in rough rice kernels in the x-ray images using the Python programming language coupled with the OpenCV library. This algorithm successfully segmented individual rice kernels in the x-ray images using the gap-filling method. The algorithm detected fissures by adaptive thresholding of each rice kernel and applying a series of filters. Data on kernel parameters (number, area, length, and width) and fissure parameters (percentage of kernels fissured and fissure number, area, and length per kernel) were produced for the images to characterize kernel size and fissuring levels of the rice sample. This algorithm demonstrated good repeatability in measuring kernel and fissure parameters, with relative standard deviations of less than 4% and 9%, respectively. The accuracy of the developed algorithm in measuring fissures was validated by visual inspection of rough rice, with a deviation of less than 2% in percentage of kernels fissured. The fissure detection and measurement algorithm provides a useful tool for quantifying fissures in rough rice samples using x-ray imaging. This information could be used to quantify fissuring levels and predict head rice yield for rough rice samples without a cumbersome milling process.