Automatic X-ray quarantine scanner and pest infestation detector for agricultural products
Highlights
► A new automatic and effective fruit quarantine system is presented. ► Guava, pitaya, pear, apple, peach, mango and Sunkist are used for experiments. ► Optimal settings for X-ray tube are presented.
Introduction
With increasing trend of international trading, large amount of agricultural products are rapidly imported or exported across international borders. The issues about how to promote the market competition, enhance the quality of products, and ensure environmental security have drawn much attention. The quarantine services of agricultural products across borders have been strengthened in many countries to reduce the possibility of invasion of alien pests, especially since the World Trade Organization (WTO) initiated the era of worldwide international market opening. Alien pests lurked in the importing agricultural products might cause extraordinary damage to local agriculture if quarantine inspection at international borders does not function well. Therefore, quarantine is a vital issue which cannot be overlooked. Several cases taught us lessons how the invasive pest caused the irreversible disaster to the sanitary region, such as the cases of Mediterranean fruit fly (Komitopoulou et al., 2004) and Asian longhorn beetle (Poland et al., 1998).
Although the importance of such service at the international borders has been recognized for a long time, most of the quarantine methods in practice still heavily depend on visual inspection. Recently, it was proven that X-ray imaging technology provided an efficient and non-destructive way for inspection of agricultural products. However, inspector’s carefully visual examination is still needed for checking the acquired X-ray images, which is very laborious and time-consuming (Kim and Schatzki, 2000, Shahin et al., 2002a, Shahin et al., 2002b). Nevertheless, checking a small portion of products is not a reliable approach since it also takes risk on chances of alien pest invasion. In order to automate the procedure of inspection of agricultural products, this work incorporated modern X-ray imaging technique with automatic diagnosis software, and we are now able to enhance the automation for the quarantine services in terms of sensitivity, specificity, accuracy, and efficiency.
Various aspects of researches on the application of X-ray inspection or quality detection have been previously reported in the field of agriculture. For example, the watercore of apple, an internal disorder that leads to tissue breakdown, can be detected in digital X-ray images (Kim and Schatzki, 2000). Shahin et al. (2002b) demonstrated that apple bruise was detectable using X-ray imaging and the extracted image features can be used to sort defected apples. In another study, they also applied the same approach to line-scanned X-ray images of sweet onions and showed that an overall classification accuracy of 90% was achieved (Shahin et al., 2002b). X-ray imaging has also shown promising results for detecting internal defects in grains or seeds. The wheat infested by weevil could be identified using the X-ray imaging technique (Karunakaran et al., 2003, Haff and Slaughter, 2004). Karunakaran et al. (2004) proposed a method of measuring the mass of wheat by calculating the total grey value from the X-ray image of the wheat. There were also several papers reporting on the applications of X-ray methods to the control of seed quality (Fesus, 1972, Singh, 1975). For example, Singh (1975) used soft X-ray to evaluate the quality of several plant seeds. The soft X-ray method is a non-destructive method that can detect insect infestation in grain kernels (Schatzki and Fine, 1988, Haff and Slaughter, 1999). Several literatures have also reported that X-ray imaging is a leading technique to detect internal defects of agricultural products (Tollner et al., 1992, Schatzki et al., 1997, Barcelon et al., 1999, Tollner, 2002, Jiang et al., 2008).
X-ray imaging technique was also used widely in quality inspection of food industry. In a series of researches for the inspection of deboned chicken meat, several image processing algorithms have been developed for the processing of X-ray images of chicken meat. Tao and Ibarra (1999) demonstrated the need for the improved methods to detect bone fragments in deboned chicken fillet of uneven thickness and proposed an image processing algorithm to eliminate the false patterns and thus enhanced the sensitivity of X-ray in bone fragment detection. Chen et al. (2003) proposed a multi-resolution-analysis-based local contrast transform to enhance the local structures of the X-ray image of chicken meat. Tao et al. (2001) proposed a method with the threshold function resulting from the smoothing of X-ray image of meat fillet. The threshold functions associated with the adaptive segmentation process were obtained by local averaging with a window whose size was dependent on the size of object to be detected. Chen et al. (2000) used a synergic laser thickness detection to adjust the grey value of the X-ray image and successfully ruled out the effect from the uneven thickness of chicken meat.
X-ray image processing techniques used in the above-mentioned applications to food industry are quite different from the ones in quarantine practice for inspecting fruit pests since the properties and constitutions of examined objects are totally different. To cope with this problem, it is essential to design an automatic and effective quarantine system equipped with suitable sensing devices and signal processing algorithms so that easy inspection procedure and high quarantine accuracy can be achieved (Yang et al., 2006). For the development of an X-ray scanning system that serves for the quarantine inspection of fruits, the aim is to process and analyze the acquired X-ray image that yields information to help quarantine inspectors in identifying possible pest infestation of the examined fruits. To achieve this, the first step to identify internal infestation of fruits by X-ray imaging technique is the image segmentation procedure to locate the infestation site. Since the grey level of X-ray images depends greatly on the density and thickness of the test samples, the relative contrast of infestation site to the intact region within a typical fruit varies with its position. Therefore, conventional automatic thresholding algorithms such as moment-preserving algorithm are not directly applicable to segment the infestation site. To solve this problem, an adaptive thresholding approach is necessary.
This research utilized X-ray images to detect internal pests of fruits. The objective of the research is to develop a prototype scanning system for automatic fruit quarantine by integrating hardware, software, and analytical algorithm components based on LabVIEW platform, and to test the feasibility of the system by various types of infected fruits. It is also desirable that optimum operating parameters of the X-ray image acquisition for different kinds of fruits or pest may be obtained through these tests. Moreover, parameters of the X-ray image analytical algorithm are obtained according to the image size. Therefore, no manual assistance is required during the process of quarantine inspection.
The rest of this paper is organized as follows. Section 2 presents the design and implementation of the X-ray quarantine scanner implemented to acquire real X-ray images of fruits. An automatic pest infestation detector is presented in Section 3. Experimental results yielded by the proposed X-ray inspection system are demonstrated in Section 4. Conclusions and discussions are given in the last section.
Section snippets
Design and implementation of X-ray quarantine scanner system
In this work, we use an X-ray scanner to acquire the image of internal tissues of fruits. By checking the acquired X-ray images, inspectors can inspect damages in the internal tissues of the fruits that are caused by external collision, or pest infestation. In order to present the proposed system systematically, this work divides the design and implementation of X-ray quarantine scanner system into three levels, including (1) user-interface level, (2) software level, and (3) hardware level. The
Pest infestation sites detector: adaptive thresholding algorithm
The intensity of the X-ray radiation transmitted through the inspected fruit depends upon the incident energy, atomic number, absorption/attenuation coefficient, density and thickness of fruit. An important characteristic of the acquired X-ray image is that its grey level of a pixel depends on the density and thickness of the inspected fruit. Due to spherical or oval shape of fruits, when X-ray passes through various thicknesses of the inspected fruits, the grey levels of infestation site in
Sample preparation
Simulation of fruit pest infestation was conducted using oriental fruit fly, Bactrocera dorsalis (Hendel), which is a serious local pest in Taiwan and there are many other species of foreign fruit flies under quarantine list with similar damage pattern on agricultural commodities. Preparation of infested fruits followed the methods of Yang et al. (2006). Source of flies were provided by the Insect Physiology and Biochemistry Laboratory of the Department of Entomology, National Chung Hsing
Conclusions
An automatic X-ray quarantine scanner system that is feasible for inspecting internal and external defects of fruits due to various pests has been developed in this work. The prototype quarantine scanner system has successfully incorporated X-ray and visible-light imaging techniques to a modular mechatronic system with precise functions of motion control, signal processing, and machine vision. The system is based on the LabVIEW development platform that offers compatibility of hardware and
Acknowledgements
The authors are grateful to the Bureau of Animal and Plant Health Inspection and Quarantine (BAPHIQ), Council of Agriculture, Taiwan, Republic of China for financially supporting this research under Grant No. 94AS-6.1.5-BQ-B1 and 95AS-7.1.5-BQ-B1. We thank K.H. Lu, National Chung Hsing University (NCHU), for providing experimental insects; T.W. Chen for comments and support of this project; China Medical University Hospital and Veterinary Medical Teaching Hospital, NCHU, for X-ray images taking
References (27)
- et al.
Robust multiple objects tracking using image segmentation and trajectory estimation scheme in video frames
Image Vis. Comput.
(2006) - et al.
An adaptive image segmentation algorithm for X-ray quarantine inspection of selected fruits
Comp. Electron. Agric.
(2008) - et al.
Medfly promoters relevant to the sterile insect technique
Insect Biochem. Mol. Biol.
(2004) - et al.
X-ray CT imaging and quality detection of peach at different physiological maturity
Trans. ASAE
(1999) - et al.
Calibration procedures for charge-coupled device X-ray detectors
Rev. Sci. Instrum.
(1999) - Chen, X., Jing, H., Tao, Y., Carr, L., Wheaton, F., 2003. Real-time Detection of Physical Hazards in Deboned Poultry...
- et al.
Multiresolution local contrast enhancement of X-ray images for poultry meat inspection
Appl. Opt.
(2000) Studies on the improvement of mass rearing for oriental fruit flies
Plant Protect. Bull.
(1978)Detection and estimation of internal pest infestation in seeds by the application of X-ray techniques
Bull. Organis. Eur. Mediter. Protect. Plantes.
(1972)- et al.
Digital Image Processing 2/e
(1993)
Real-time X-ray inspection of wheat for infestation by the granary weevil, Sitophilus granarius (L.)
Trans. ASAE
Cited by (28)
A smartphone application for site-specific pest management based on deep learning and spatial interpolation
2024, Computers and Electronics in AgricultureAutomated detection of internal fruit rot in Hass avocado via deep learning-based semantic segmentation of X-ray images
2023, Postharvest Biology and TechnologyDevelopment of automatic detection model for stem-end rots of ‘Hass’ avocado fruit using X-ray imaging and image processing
2022, Postharvest Biology and TechnologyCitation Excerpt :X-rays are advantageous for internal quality measurement (especially for fruit with thick skin) because X-rays penetrate deep inside an object (Arendse et al., 2018). Since the 2000 s, various image segmentation methods for the X-ray line-scanned images of agricultural products have been developed (Chuang et al., 2011; Dael et al., 2016; Jiang et al., 2008; Mathanker et al., 2010). Ogawa et al. (2003) reported that X-ray line-scanned images can reveal “split-pits,” a problem in the cores of apples and peaches.
Emerging non-destructive imaging techniques for fruit damage detection: Image processing and analysis
2022, Trends in Food Science and TechnologyMachine learning applications to non-destructive defect detection in horticultural products
2020, Biosystems EngineeringInline discrete tomography system: Application to agricultural product inspection
2017, Computers and Electronics in AgricultureCitation Excerpt :Conventional X-ray radiography has been extensively used in inspection and quality assurance of agricultural products on a conveyor belt. A typical example is food inspection for detecting defects or foreign objects (e.g., insect pests, stones, glass, metal parts) (Ayalew et al., 2004; Jiang et al., 2008; Chuang et al., 2011; Mery et al., 2011). While X-ray radiographs provide some information about the internal structure of objects, one of the major disadvantages of traditional radiography is that it cannot provide quantitative 3D information about the object to be scanned.
- 1
Equal contributors.