Study of the Accuracy of Apple Internal Lesion Detection Based on Frequency Domain Diffuse Optical Tomography
LI Jiang-tao1, HU Wen-yan1, ZHAO Long-lian1, 2*, LI Jun-hui1, 2
1. College of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China
2. Modern Precision Agriculture System Integration Research Key Laboratory, Ministry of Education, Beijing 100083, China
Abstract：Lesions inside the apple tissue cause changes in their optical parameters. Frequency domain diffuse optical tomography（FD-DOT） is used to detect the absorption coefficient and the reduced scattering coefficient of apple tissue, and the reconstructed image obtained by combining the 3D reconstruction technology can intuitively understand the internal condition of the apple. In this way, non-destructive detection of internal lesions in apple was realized. According to the visible near-infrared transmission spectrum of normal and rotten apple samples, the wavelength of 740nm can be chosen as the laser light source to distinguish normal and diseased apple tissues furthest. The imaging precision will vary with the frequency of incident light modulation, the degree, location, and size of lesions within the apple. In this paper, a series of simulation experiments are designed to study the effect of the above factors on the detection accuracy: to study the influence of modulation frequency on the accuracy of reconstructed images, setting different laser modulation frequencies; to study the influence of the size of lesion on the accuracy of the reconstructed image, adding spherical heteroplasmon of different sizes at a certain position in the apple model ; A heteroplasmon of a certain size was added at different positions to study the influence of different lesion positions on the accuracy of the reconstructed image. Firstly, finite element mesh model of apple was established by Abaqus. Twelve 740nm near-infrared laser sources and six detectors were designed to be evenly arranged on the surface of the apple model. Then according to the experimental needs, a spherical heteroplasmon representing lesion was added to the tissue model. After that irradiation into interior of the apple with a high-frequency modulated light source, and detect the AC amplitude and phase delay of the emitted light. The software NIRFAST is used to calculate and inversely derive the absorption coefficient and the reduced scattering coefficient distribution of the apple to be tested. Finally, the reconstructed image is obtained by using 3D reconstruction method, reconstruction results can be evaluated using the absorption coefficient contrast-to-noise ratio (CNR) and absorption coefficient distribution of the reconstructed image. The experimental results show that in order to detect deep lesions of larger apples, a higher incident light modulation frequency is required; this method can detect most spherical lesions with a radius greater than 5 mm in suitable apples, and as the size of lesions is enlarged within a certain range, the accuracy of the reconstructed image is gradually increased. However, when lesion area is too large, the image accuracy begins to decrease. When lesion is closer and closer to the detector, the accuracy of the reconstructed image gradually increases, but when distance between lesion and detector is too small, the accuracy of the reconstructed image has a tendency to decrease; the closer the vertical distance of lesion region to the detector plane is, the higher the accuracy of the reconstructed image is. The above experimental results will lay a good foundation for the application of FD-DOT in non-destructive testing of apples.
李江涛，胡文雁，赵龙莲 ，李军会. 频域近红外光学成像法的苹果内部病变检测精度[J]. 光谱学与光谱分析, 2019, 39(09): 2836-2841.
LI Jiang-tao, HU Wen-yan, ZHAO Long-lian, LI Jun-hui. Study of the Accuracy of Apple Internal Lesion Detection Based on Frequency Domain Diffuse Optical Tomography. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2019, 39(09): 2836-2841.
[1] Folch-Fortuny A, Prats-Montalbán J M, Cubero S, et al. Chemometrics & Intelligent Laboratory Systems, 2016, 156: 241.
[2] Liu Yande, Zhou Yanrui. Spectroscopy，2013, 28(7): 32.
[3] Qin J, Lu R, Peng Y. Proc. SPIE, 2007, 6761: 67610M.
[4] Lu R, Cen H, Huang M. Transactions of the ASABE, 2010, 53(1): 263.
[5] SUN Mei-xiu，KANG Mei-ling，LI Ying-xin, et al（孙美秀，康美玲，李迎新，等）. International Journal of Biomedical Engineering（国际生物医学工程杂志）, 2012, 35（1）: 14.
[6] Li X D，Yodh A G, Chanee B. Optics Letters, 1997, 22（8）: 573.
[7] Whalen C, Maclin E L, Fabiani M, et al. Human Brain Mapping，2008, 29(11): 1288.
[8] Yamada Y, Okawa S. Optical Review, 2014, 21(3): 185.
[9] Dehghani H, Pogue B W, Poplack S P, et al. Applied Optics, 2003, 42: 135.
[10] Chance B, Luo Q M, Nioka S, et al. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 1997, 352(1354): 707.
[11] Song Xiaomei, Brian W Pogue，Jiang Shudong, et al. Applied Optics，2004, 43(5): 1053.
[12] GUO Zhi-ming, HUANG Wen-qian, CHEN Quan-sheng, et al(郭志明，黄文倩，陈全胜，等）. Transactions of the Chinese Society of Agricultural Engineering (农业工程学报）. 2016, 32(6): 283.
[13] Dehghani H, Eames M E, Yalavarthy P K, et al. Communications In Numerical Methods in Engineering, 2008, (6): 711.
