Infrared image super-resolution using auxiliary convolutional neural network and visible image under low-light conditions

Han, Tae Young,Kim, Dae Ha,Lee, Seung Hyun,Song, Byung Cheol Elsevier 2018 Journal of visual communication and image represen Vol.51 No.-

<P><B>Abstract</B></P> <P>Convolutional neural networks (CNN) have been successfully applied to visible image super-resolution (SR) methods. In this study, we propose a CNN-based SR algorithm for up-scaling near-infrared (NIR) images under low-light conditions, using corresponding visible images. Our algorithm first extracts high-frequency (HF) components from the up-scaled low-resolution (LR) NIR image and its corresponding high-resolution (HR) visible image, and then takes them as multiple inputs of the CNN. Next, the CNN outputs the HR HF component of the input NIR image. Finally, an HR NIR image is synthesized by adding the HR HF component to the up-scaled LR NIR image. The simulation results show that the proposed algorithm outperforms the state-of-the-art methods, in terms of both qualitative and quantitative aspects.</P> <P><B>Highlights</B></P> <P> <UL> <LI> The proposed scheme focuses on SR of an NIR image, not a VIS image in a low-light environment. </LI> <LI> The proposed scheme utilizes a VIS image, obtained at the same time as the NIR image, as auxiliary information. </LI> <LI> The proposed scheme is based on a CNN structure that simultaneously receives the HF information of the NIR and VIS images. </LI> <LI> The proposed scheme achieved a further performance improvement by employing an auxiliary CNN network. </LI> </UL> </P>

가시광선, 근적외선, 자외선에 의한 혀의 영상 특성 비교연구

이상석,남동현,노재혁,이충희 한국물리학회 2011 새물리 Vol.61 No.5

In this study, we developed a medical tongue diagnosis device using three wavelengths of light: visible, near infrared (940 nm) and ultraviolet (375 nm). The device is composed of a light source equipped with visible, near infrared, and ultraviolet light-emitting diodes (LEDs), a photography section using a hivision camera, and a control section to obtain three tongue images. The near-infrared tongue image for imaging vein blood below the tongue showed the feature of tongue vein blood. One of the ultraviolet tongue images exhibited three fluorescent colors: red, green, and blue. The moisture on the tongue's surface, the thickness of a coating on the tongue, and the degree of fluorescent color could be established by comparing several tongue images: a wet tongue, a coated tongue, and a bongur painted with water, mayonez, and fluorescence pen color. 혀의 영상을 촬영하기 위한 장치인 설진기를 3파장대 광원인 가시광선,근적외선, 자외선 발광다이오드(LED)를 이용하여 제작하였다. 설진기는가시광선, 근적외선, 자외선 LED를 사용한 광원부와 반도체 소자가장착된 하이비전 카메라인 촬상부 및 가시광선, 근적외선, 자외선 영상을얻기 위한 제어부로 구성되어 있다. 근적외선 촬영은 설하정맥 영역의분리에 유리하며 자외선 형광 촬영은 설태의 종류를 구분하는데활용하였다. 혀에 물, 마요네즈, 형광펜을 독립적으로 묻혀 촬영한설진영상들의 비교를 통해 근적외선 영상으로 습윤도의 정보 획득과자외선 영상을 통해 황 화합물이 주성분인 설태의 두께와 형광정도의구분이 가능하였다.

웨이블릿 변환 영역의 부대역 정보를 이용한 가시광 및 근적외선 영상 합성

이현우(Hyeon-U Lee),권혁주(Hyuk-Ju Kwon),이성학(Sung-Hak Lee) 한국정보기술학회 2020 Proceedings of KIIT Conference Vol.2020 No.10

본 논문에서는 가시광 및 근적외선 영상을 촬영하고 Cohen-Daubechies–Feauveau 9/7을 이용한 웨이브릿(wavelet) 변환영역에서 부대역(subband) 에너지 분포를 이용하여 두 영상의 성분을 분리 합성하는 방법을 제시한다. 제안한 영상 합성 방법은 영상 부대역 변환, 계층 분할, 다중 대역 합성 후 색채 보상 단계로 이루어지고, 가시광 및 근적외선의 다중 대역 영상 합성을 위한 부대역 가중치 맵을 제안한다. 제안한 방법은 실험결과에서 자연스러운 경계 표현에서 기존의 방법보다 우수함을 나타내었다. 또한 특수 센서를 사용하지 않고 side by side 듀얼 카메라를 사용하여 실시간 영상 촬영 및 합성이 가능하다는 점에서 빠른 처리 속도와 활용도 면에서 강점을 보인다. In this paper, the method of taking visible light and near-infrared images and separating and synthesizing the components of images using the subband energy distribution of Cohen-Dauvechies–Feauveau 9/7 wavelet transform is presented. The proposed image synthesis method consists of wavelet subband transform, layer separation, multi-layer image synthesis, and color compensation stages. Moreover the subband weight map is proposed for image synthesis of visible light and near-infrared multiple bands. The proposed method indicates that it is superior to the existing method in natural boundary expression. In addition, it shows its strength in fast processing speed and utilization in that it is possible to shoot and synthesize images in real time using side by side dual cameras without using special sensors.

웨이블릿 변환 영역의 부대역 정보를 이용한 가시광 및 근적외선 영상 합성

이현우(Hyeon-U Lee),권혁주(Hyuk-Ju Kwon),이성학(Sung-Hak Lee) 한국정보기술학회 2021 한국정보기술학회논문지 Vol.19 No.3

In this paper, we propose the method of taking visible light and near-infrared images and synthesizing the components of images using the subband energy distribution of Cohen-Dauvechies–Feauveau 9/7 wavelet transform. The proposed image synthesis method consists of wavelet subband transform, layer separation, multi-layer image synthesis, and color compensation stages. Moreover the subband weight map is proposed for image synthesis of visible light and near-infrared multiple bands. It is shown that the proposed method is superior to the existing method in natural boundary expression. In addition, it shows its strength in fast processing speed and utilization in that it is possible to shoot and synthesize images in real time using side by side dual cameras without using special sensors.

결합형 양방향 필터를 이용한 주야간 적응형 가시광 및 근적외선 영상 합성

박철우,권혁주,이성학 한국정보기술학회 2022 한국정보기술학회논문지 Vol.20 No.6

In this paper, we present a broadband imaging technique to identify distant objects in day and night time. In order to synthesize the two acquired visible and near-infrared images, first, a base component and a detail component are decomposed using a bilateral filter, and the detailed components are synthesized using a local variance map. did Also, considering the difference in NIR imaging characteristics in daytime and nighttime modes, the basis component was synthesized using different methods of luminance saturation map using joint bilateral filer and depth and penetration map. The proposed method overcomes the partial overexposure or underexposure phenomenon caused by strong sun and IR lights used in broadband imaging and improves the identification of objects in various spaces than the existing method by emphasizing detailed components. 본 논문에서는 주야간의 사물 식별을 위한 광대역 영상 합성 기법을 제시한다. 취득된 가시광 및 근적외선 두 영상을 합성하기 위해 먼저, 양방향 필터(Bilateral filter)를 이용하여 기저 성분과 세부 성분으로 분해하고 국부 분산맵을 이용하여 세부 성분을 합성한다. 또한 주간과 야간의 근적외선 영상 특성의 차이를 고려하여 결합형 양방향 필터를 이용한 휘도 포화 영역 맵과 깊이 및 침투 맵을 이용하여 기저 성분을 합성하였다. 제안 방법은 광대역 영상 촬영에서 다양하게 경험되는 태양광과 적외선 보조광에 의한 부분적 과도 또는 부족 노출 현상을 극복하고 세부 성분을 강조하여 기존의 방법보다 실내 실외의 다양한 영상에서 사물 식별을 개선하는 결과를 보인다.

Spatial assessment of soluble solid contents on apple slices using hyperspectral imaging

Mo, C.,Kim, M.S.,Kim, G.,Lim, J.,Delwiche, S.R.,Chao, K.,Lee, H.,Cho, B.K. Elsevier Science B.V., Amsterdam 2017 BIOSYSTEMS ENGINEERING Vol.159 No.-

<P>A partial least squares regression (PLSR) model to map internal soluble solids content (SSC) of apples using visible/near-infrared (VNIR) hyperspectral imaging was developed. The reflectance spectra of sliced apples were extracted from hyperspectral absorbance images obtained in the 400-1000 nm range. Prediction models for SSC mapping were developed for three different measurement/sampling designs that varied in the number and size of the regions of interest (ROIs) used for apple SSC measurement and spectral averaging. Case 1 used 29 small ROIs per apple, Case II used 9 moderate-size ROIs per apple, and Case III used 5 large ROIs per apple. The optimal pre-treatment of the spectra extracted from the hyperspectral images was investigated to enhance the performance of the prediction models. The coefficients of determination and root mean square errors of the best-performing models were, respectively, 0.802 and +/- 0.674 degrees Brix for Case I, 0.871 and +/- 0.524 degrees Brix for Case II, and 0.876 and +/- 0.514 degrees Brix for Case III. The accuracy of the PLSR models was enhanced by using the spectra and SSC measured/averaged from the fewer but larger areas of the apples rather than from more numerous but smaller areas. PLS images of SSC showed the predicted internal distribution of SSC within the apples. The overall results demonstrate that hyperspectral absorbance imaging techniques may be useful for mapping internal soluble solids content of apples. (C) 2017 IAgrE. Published by Elsevier Ltd. All rights reserved.</P>

샤크-하트만 센서를 이용한 가시광 및 근적외선 분광기 조립 및 평가

황성령,이준호,정도환,홍진석,김영수,김연수,김현숙 한국광학회 2017 한국광학회지 Vol.28 No.3

We report the assembly procedure and performance evaluation of a visible and near-infrared spectrometer in the wavelengthregion of 400-900 nm, which is later to be combined with fore-optics (a telescope) to form a f/2.5 imaging spectrometer witha field of view of ±7.68°. The detector at the final image plane is a 640×480 charge-coupled device with a 24 μm pixel size. The spectrometer is in an Offner relay configuration consisting of two concentric, spherical mirrors, the secondary of which isreplaced by a convex grating mirror. A double-pass test method with an interferometer is often applied in the assembly processof precision optics, but was excluded from our study due to a large residual wavefront error (WFE) in optical design of 210nm (0.35λ at 600 nm) root-mean-square (RMS). This results in a single-path test method with a Shack-Hartmann sensor. Thefinal assembly was tested to have a RMS WFE increase of less than 90 nm over the entire field of view, a keystone of 0.08pixels, a smile of 1.13 pixels and a spectral resolution of 4.32 nm. During the procedure, we confirmed the validity of usinga Shack-Hartmann wavefront sensor to monitor alignment in the assembly of an Offner-like spectrometer.

Effectiveness of Using the TIR Band in Landsat 8 Image Classification

이미희,이수봉,김용민,사지원,어양담 한국측량학회 2015 한국측량학회지 Vol.33 No.3

This paper discusses the effectiveness of using Landsat 8 TIR (Thermal Infrared) band images to improve the accuracy of landuse/landcover classification of urban areas. According to classification results for the study area using diverse band combinations, the classification accuracy using an image fusion process in which the TIR band is added to the visible and near infrared band was improved by 4.0%, compared to that using a band combination that does not consider the TIR band. For urban area landuse/landcover classification in particular, the producer’s accuracy and user’s accuracy values were improved by 10.2% and 3.8%, respectively. When MLC (Maximum Likelihood Classification), which is commonly applied to remote sensing images, was used, the TIR band images helped obtain a higher discriminant analysis in landuse/landcover classification.

야간의 사물인식 개선을 위한 가시광 및 근적외선 영상 합성

이성학(Sung-hak Lee),손동민(Dong-min Son),정태영(Tae-young Jung) 한국정보기술학회 2019 Proceedings of KIIT Conference Vol.2019 No.6

Effectiveness of Using the TIR Band in Landsat 8 Image Classification

Lee, Mi Hee,Lee, Soo Bong,Kim, Yongmin,Sa, Jiwon,Eo, Yang Dam 한국측량학회 2015 한국측량학회지 Vol.33 No.3