• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.3106-3106
• /
• 2001

Imaging spectrometry or hyperspectral imaging is a recent development that makes possible quantitative and qualitative measurement for food quality and safety. This paper presents the research results that a hyperspectral imaging system can be used effectively for detecting fecal (from duodenum, cecum, and colon) and ingesta contamination on poultry carcasses from the different feed meals (wheat, mile, and corn with soybean) for poultry safety inspection. A hyperspectral imaging system has been developed and tested for the identification of fecal and ingesta surface contamination on poultry carcasses. Hypercube image data including both spectral and spatial domains between 430 and 900 nm were acquired from poultry carcasses with fecal and ingesta contamination. A transportable hyperspectral imaging system including fiber optically fabricated line lights, motorized lens control for line scans, and hypercube image data from contaminated carcasses with different feeds are presented. Calibration method of a hyperspectral imaging system is demonstrated using different lighting sources and reflectance panels. Principal Component and Minimum Noise Fraction transformations will be discussed to characterize hyperspectral images and further image processing algorithms such as image band ratio of dual-wavelength images and its histogram stretching with thresholding process will be demonstrated to identify fecal and ingesta materials on poultry carcasses. This algorithm could be further applied for real-time classification of fecal and ingesta contamination on poultry carcasses in the poultry processing line.

• Korean Journal of Agricultural Science
• /
• v.47 no.3
• /
• pp.645-655
• /
• 2020

Bacteria are a very common cause of food poisoning. Moreover, bacteria form biofilms to protect themselves from harsh environments. Conventional detection methods for foodborne bacterial pathogens including the plate count method, enzyme-linked immunosorbent assays (ELISA), and polymerase chain reaction (PCR) assays require a lot of time and effort. Hyperspectral imaging has been used for food safety because of its non-destructive and real-time detection capability. This study assessed the feasibility of using hyperspectral imaging and machine learning techniques to detect biofilms formed by Escherichia coli. E. coli was cultured on a high-density polyethylene (HDPE) coupon, which is a main material of food processing facilities. Hyperspectral fluorescence images were acquired from 420 to 730 nm and analyzed by a single wavelength method and machine learning techniques to determine whether an E. coli culture was present. The prediction accuracy of a biofilm by the single wavelength method was 84.69%. The prediction accuracy by the machine learning techniques were 87.49, 91.16, 86.61, and 86.80% for decision tree (DT), k-nearest neighbor (k-NN), linear discriminant analysis (LDA), and partial least squares-discriminant analysis (PLS-DA), respectively. This result shows the possibility of using machine learning techniques, especially the k-NN model, to effectively detect bacterial pathogens and confirm food poisoning through hyperspectral images.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.43 no.4 s.310
• /
• pp.11-20
• /
• 2006

This paper presents a method for detecting tumors on chicken carcasses by fusion of hyperspectral fluorescence and reflectance images. Classification of normal skin and tumor is performed by the image obtain 어 from optimal band ratio which minimizes the overlapping area of PDFs for normal skin and tumor. This method yields four feature images, each of them represents the ratio of two intensity values from a pixel. Classification is achieved by applying ISODATA to each pixel from the feature images. For the analysis of reflectance image, band selection method is proposed based on the information quantity, many effective features are acquired for the classification by defining the linear transformation selecting the projection axis, accordingly, accurate interpretation of images is possible in the reflectance image and automatic feature selection method is realized. Feature images from reflectance images are also classified by ISODATA and combined with the result from fluorescence images. Experimental result indicates that improved performance in term of reducing false detection rate is observed.

• Journal of Biosystems Engineering
• /
• v.38 no.3
• /
• pp.199-207
• /
• 2013

Purpose: Tomatoes, an important agricultural product in fresh-cut markets, are sometimes a source of foodborne illness, mainly Salmonella spp. Growth cracks on tomatoes can be a pathway for bacteria, so its detection prior to consumption is important for public health. In this study, multispectral Visible/Near-Infrared (NIR) reflectance imaging techniques were used to determine optimal wavebands for the classification of defect tomatoes. Methods: Hyperspectral reflectance images were collected from samples of naturally cracked tomatoes. To classify the resulting images, the selected wavelength bands were subjected to two-band permutations, and a supervised classification method was used. Results: The results showed that two optimal wavelengths, 713.8 nm and 718.6 nm, could be used to identify cracked spots on tomato surfaces with a correct classification rate of 91.1%. The result indicates that multispectral reflectance imaging with optimized wavebands from hyperspectral images is an effective technique for the classification of defective tomatoes. Conclusions: Although it can be susceptible to specular interference, the multispectral reflectance imaging is an appropriate method for commercial applications because it is faster and much less expensive than Near-Infrared or fluorescence imaging techniques.

• Journal of the Optical Society of Korea
• /
• v.20 no.6
• /
• pp.752-761
• /
• 2016

Hyperspectral images are often contaminated with stripe noise, which severely degrades the imaging quality and the precision of the subsequent processing. In this paper, a variational model is proposed by employing spectral-spatial adaptive unidirectional variation and a sparse representation. Unlike traditional methods, we exploit the spectral correction and remove stripes in different bands and different regions adaptively, instead of selecting parameters band by band. The regularization strength adapts to the spectrally varying stripe intensities and the spatially varying texture information. Spectral correlation is exploited via dictionary learning in the sparse representation framework to prevent spectral distortion. Moreover, the minimization problem, which contains two unsmooth and inseparable $l_1$-norm terms, is optimized by the split Bregman approach. Experimental results, on datasets from several imaging systems, demonstrate that the proposed method can remove stripe noise effectively and adaptively, as well as preserve original detail information.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.24 no.1
• /
• pp.12-25
• /
• 2021

Land cover maps are used in various fields in urban expansion and development. This study analyzed the amount of land cover change over time using multi-sensor information, focusing on the waterfront area of the Taehwa River. In order to apply high-accuracy aerial hyperspectral images, patterns with Field-spectral were reviewed and compared with time series Digital map. The hyperspectral image was set as 13 land cover grades, and the time series digital map was classified into 7 and the waterfront area was classified into 5-6 grades and analyzed. As a result of analysis of the change in land cover of the digital map from the 1990s to 2010, it was found that forest areas were rapidly decreasing and Farmland and grassland were becoming urban. As for the land cover change(2010~2019) in the waterfront area(set 500m) analyzed through hyperspectral images, it was found that Farmland(1.4㎢), Forest(1.0㎢), and grassland (0.8㎢) were converted into urbanized and dried areas, and urbanization was accelerating around the Taehwa River waterfront. Recently, a lot of research has been conducted on the production of land cover maps using high-precision satellite images and aerial hyperspectral images, so it is expected that more detailed and precise land cover maps can be produced and utilized.

• Journal of Korea Water Resources Association
• /
• v.55 no.4
• /
• pp.301-308
• /
• 2022

Algae is an indispensable primary producer in the ecosystem by supplying energy to consumers in the aquatic ecosystem, and is largely divided into green algae, blue-green algae, and diatoms. In the case of blue-green algae, the water temperature rises, which occurs in the summer and overgrows, which is the main cause of the algae bloom. Recently, the change in the occurrence time and frequency of the algae bloom is increasing due to climate change. Existing algae survey methods are performed by collecting water and measuring through sensors, and time, cost and manpower are limited. In order to overcome the limitations of these existing monitoring methods, research has been conducted to perform remote monitoring using spectroscopic devices such as multispectral and hyperspectral using satellite image, UAV, etc. In this study, we tried to confirm the possibility of species classification of remote monitoring through laboratory-scale experiments through algal culture and river water collection. In order to acquire hyperspectral images, a hyperspectral sensor capable of analyzing at 400-1000 nm was used. In order to extract the spectral characteristics of the collected river water for classification of algae species, filtration was performed using a GF/C filter to prepare a sample and images were collected. Radiation correction and base removal of the collected images were performed, and spectral information for each sample was extracted and analyzed through the process of extracting spectral information of algae to identify and compare and analyze the spectral characteristics of algae, and remote sensing based on hyperspectral images in rivers and lakes. We tried to review the applicability of monitoring.

• Structural Monitoring and Maintenance
• /
• v.2 no.3
• /
• pp.181-197
• /
• 2015

Assessing the condition of paint on civil structures is an important but challenging and costly task, in particular when it comes to large and complex structures. Current practices of visual inspection are labour-intensive and time-consuming to perform. In addition, this task usually relies on the experience and subjective judgment of individual inspectors. In this study, hyperspectral imaging and classification techniques are proposed as a method to objectively assess the state of the paint on a civil or other structure. The ultimate objective of the work is to develop a technology that can provide precise and automatic grading of paint condition and assessment of degradation due to age or environmental factors. Towards this goal, we acquired hyperspectral images of steel surfaces located at long (mid-range) and short distances on the Sydney Harbour Bridge with an Acousto-Optics Tunable filter (AOTF) hyperspectral camera (consisting of 21 bands in the visible spectrum). We trained a multi-class Support Vector Machines (SVM) classifier to automatically assess the grading of the paint from hyperspectral signatures. Our results demonstrate that the classifier generates highly accurate assessment of the paint condition in comparison to the judgement of human experts.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.22 no.4
• /
• pp.1-11
• /
• 2019

With the development of remote sensing sensor technology, it has become possible to acquire satellite images with various spectral information. In particular, since the hyperspectral image is composed of continuous and narrow spectral wavelength, it can be effectively used in various fields such as land cover classification, target detection, and environment monitoring. Change detection techniques using remote sensing data are generally performed through differences of data with same dimensions. Therefore, it has a disadvantage that it is difficult to apply to heterogeneous sensors having different dimensions. In this study, we have developed a change detection method applicable to hyperspectral image and high spat ial resolution satellite image with different dimensions, and confirmed the applicability of the change detection method between heterogeneous images. For the application of the change detection method, the dimension of hyperspectral image was reduced by using correlation analysis and principal component analysis, and the change detection algorithm used CVA. The ROC curve and the AUC were calculated using the reference data for the evaluation of change detection performance. Experimental results show that the change detection performance is higher when using the image generated by adequate dimensionality reduction than the case using the original hyperspectral image.

• Korean Journal of Remote Sensing
• /
• v.35 no.6_1
• /
• pp.959-971
• /
• 2019

CNN (Convolutional Neural Network) is one representative deep learning algorithm, which can extract high-level spatial and spectral features, and has been applied for hyperspectral image classification. However, one significant drawback behind the application of CNNs in hyperspectral images is the high dimensionality of the data, which increases the training time and processing complexity. To address this problem, several CNN based hyperspectral image classification studies have exploited PCA (Principal Component Analysis) for dimensionality reduction. One limitation to this is that the spectral information of the original image can be lost through PCA. Although it is clear that the use of PCA affects the accuracy and the CNN training time, the impact of PCA for CNN based hyperspectral image classification has been understudied. The purpose of this study is to analyze the quantitative effect of PCA in CNN for hyperspectral image classification. The hyperspectral images were first transformed through PCA and applied into the CNN model by varying the size of the reduced dimensionality. In addition, 2D-CNN and 3D-CNN frameworks were applied to analyze the sensitivity of the PCA with respect to the convolution kernel in the model. Experimental results were evaluated based on classification accuracy, learning time, variance ratio, and training process. The size of the reduced dimensionality was the most efficient when the explained variance ratio recorded 99.7%~99.8%. Since the 3D kernel had higher classification accuracy in the original-CNN than the PCA-CNN in comparison to the 2D-CNN, the results revealed that the dimensionality reduction was relatively less effective in 3D kernel.