X-ray computed tomography for quality inspection of agricultural products: A review

Abstract

The quality of agricultural products relates to the internal structure, which has long been a matter of interest in agricultural scientists. However, inspection methods of the opaque nature of internal information on agricultural products are usually destructive and require sample separation or preparation. X-ray computed tomography (X-ray CT) technology is one of the important nondestructive testing (NDT) technologies without sample separation and preparation. In this study, X-ray CT technology is used to obtain two-dimensional slice images and three-dimensional tomographic images of samples. The purpose of the review was to provide an overview of the working principle of X-ray CT technology, image processing, and analysis. This review aims to focus on the development of the agricultural products (e.g., wheat, maize, rice, apple, beef) and its applications (e.g., internal quality evaluation, microstructure observation, mechanical property measurement, and others) using CT scanner. This paper covers the aspects regarding the advantages and disadvantages of NDT technology, especially the unique advantages and limitations of X-ray CT technology on the quality inspection of agricultural products. Future prospects of X-ray CT technology are also put forward to become indispensable to the quality evaluation and product development on agricultural products.

Keywords: agricultural products; computed tomography; nondestructive; quality inspection.

Figure 1

Working principle of X‐ray CT technology. Some radial projections are captured at different angles on a sample by the X‐ray to obtain 2D slice images. Besides, a 3D image could be produced by the restructure of scanned numerous slice images of the sample

Figure 2

Image processing and analysis procedure. 2D slice images are generated to reconstruct the 3D model. Gaussian or median filter is used to reduce noise with 3D raw gray‐level images. The threshold segmentation is used to segment the image based on the gray value histogram of different regions. The final step is the qualitative and quantitative analysis on CT data of ROIs

Figure 3

3D visualizations of wheat grain (Suresh & Neethirajan, 2015). (a) Vertical cross section of insect‐infected wheat grain; (b) horizontal cross section of sprout‐damaged wheat grain; (c) region of interest in the insect‐infected wheat grain alongside its ortho‐projection

Figure 4

Spatial distribution of density within kernels by X‐ray CT. WTR is wild‐type rice; HAR is high‐amylose rice (Zhu et al., 2012). The instruments are SEM with an accelerating voltage of 30 kV and a high‐resolution X‐ray CT system with operating voltage of 46 kV and current of 75 μA. The amount of void was the least near the tip part of the WTR grain and varied throughout grain. Within the HAR grain, the part near the tip also had the least amount of void, but no difference was found in other parts of the grain

Figure 5

Histogram of durum wheat grain transmitted light images. (a) Vitreous; (b) nonvitreous. (0 represents black and 255 represents white in the x‐axis) (Neethirajan et al., 2006). The images were acquired by X‐ray CT at 17 kV potential, 65 μA current, and a resolution of 60 pixels/mm. Nonvitreous kernels have more optically dense regions than the translucent vitreous kernels

Figure 6

Comparison of a longitudinal digital image and CT image slice. (a) A longitudinal digital image; (b) 2D CT image slice (Guelpa et al., 2015). The voxel size, voltage, and scan time are 13.4 μm, 60 kV, and 30 min, respectively. The same maize grain is depicting the internal structure of the maize grain, that is, flour and vitreous endosperm, germ, and pedicle. In CT image, the brighter gray region represents the denser vitreous endosperm and the darker region the less dense floury endosperm. The vitreous endosperm thus appeared translucent (Figure a) due to no light being reflected