• 센서학회지
• /
• 제19권2호
• /
• pp.118-123
• /
• 2010

This paper proposes super-resolution reconstruction algorithm for image enhancement. Super-resolution reconstruction algorithms reconstruct a high-resolution image from multi-frame low-resolution images of a scene. Conventional super- resolution reconstruction algorithms are iterative back-projection(IBP), robust super-resolution(RS)method and standard Papoulis-Gerchberg(PG)method. However, traditional methods have some problems such as rotation and ringing. So, this paper proposes modified algorithm to improve the problem. Experimental results show that this proposed algorithm solve the problem. As a result, the proposed method showed an increase in the PSNR for traditional super-resolution reconstruction algorithms.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2009년도 IWAIT
• /
• pp.254-257
• /
• 2009

In this paper, we propose a learning-based super-resolution algorithm. In the proposed algorithm, a multi-resolution wavelet approach is adopted to perform the synthesis of local high-frequency features. To obtain a high-resolution image, wavelet coefficients of two dominant LH- and HL-bands are estimated based on wavelet frames. In order to prepare more efficient training sets, the proposed algorithm utilizes the LH-band and transposed HL-band. The training sets are then used for the estimation of wavelet coefficients for both LH- and HL-bands. Using the estimated high frequency bands, a high resolution image is reconstructed via the wavelet transform. Experimental results demonstrate that the proposed scheme can synthesize high-quality images.

• Journal of Power Electronics
• /
• 제13권6호
• /
• pp.955-963
• /
• 2013

To acquire a performed and practical solution that is free from chattering, this study proposes the use of an adaptive super-twisting algorithm to drive a single-phase induction motor. Partial feedback linearization is applied before using a super-twisting algorithm to control the speed and stator currents. The load torque is considered an unknown but bounded disturbance. Therefore, a time-varying switching gain that does not require prior knowledge of the disturbance boundary is proposed. A simple sliding surface is formulated as the difference between the real and desired trajectories obtained from the indirect rotor flux oriented control strategy. To illustrate the effectiveness of the proposed control structure, an experimental setup around a digital signal processor (dS1104) is developed and several tests are performed.

• Nuclear Engineering and Technology
• /
• 제42권1호
• /
• pp.47-54
• /
• 2010

The SuperCritical Water-cooled Reactor (SCWR) pursues high power density to reduce its capital cost. The fast spectrum SCWR, called a super fast reactor, can be designed with a higher power density than thermal spectrum SCWR. The mechanism of increasing the average power density of the super fast reactor is studied theoretically and numerically. Some key parameters affecting the average power density, including fuel pin outer diameter, fuel pitch, power peaking factor, and the fraction of seed assemblies, are analyzed and optimized to achieve a more compact core. Based on those sensitivity analyses, a compact super fast reactor is successfully designed with an average power density of 294.8 W/$cm^3$. The core characteristics are analyzed by using three-dimensional neutronics/thermal-hydraulics coupling method. Numerical results show that all of the design criteria and goals are satisfied.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권11호
• /
• pp.4065-4083
• /
• 2021

Super-resolution can improve the clarity of low-resolution (LR) images, which can increase the accuracy of high-level compute vision tasks. Portable devices have low computing power and storage performance. Large-scale neural network super-resolution methods are not suitable for portable devices. In order to save the computational cost and the number of parameters, Lightweight image processing method can improve the processing speed of portable devices. Therefore, we propose the Enhanced Information Multiple Distillation Network (EIMDN) to adapt lower delay and cost. The EIMDN takes feedback mechanism as the framework and obtains low level features through high level features. Further, we replace the feature extraction convolution operation in Information Multiple Distillation Block (IMDB), with Ghost module, and propose the Enhanced Information Multiple Distillation Block (EIMDB) to reduce the amount of calculation and the number of parameters. Finally, coordinate attention (CA) is used at the end of IMDB and EIMDB to enhance the important information extraction from Spaces and channels. Experimental results show that our proposed can achieve convergence faster with fewer parameters and computation, compared with other lightweight super-resolution methods. Under the condition of higher peak signal-to-noise ratio (PSNR) and higher structural similarity (SSIM), the performance of network reconstruction image texture and target contour is significantly improved.

• Wind and Structures
• /
• 제21권3호
• /
• pp.331-352
• /
• 2015

A series of wind tunnel tests were conducted on tapered super high-rise buildings with a square cross section by applying synchronous pressure measurement technology. The effects of global strategy of chamfered modification on aerodynamic loads and wind-induced responses were investigated. Moreover, local aerodynamic strategies of opening a ventilation slot in the corner of equipment and refuge floors were carried out. Results show that the global strategy of tapered elevation increased the vortex shedding frequency, but reduced vortex shedding energy, leading to reduction of across-wind aerodynamic loads and responses. Chamfered modification suppressed the across-wind vortex shedding effect on tapered buildings. Opening the ventilation slot further suppressed the strength of vortex shedding and reduced the residual energy related to vortex shedding in aerodynamic loads of chamfered buildings. Finally, the optimized locations of local aerodynamic strategies were suggested.

• 한국작물학회지
• /
• 제44권3호
• /
• pp.267-272
• /
• 1999

In order to determine the optimum harvest time of vegetable corns, the changes in sugars, soluble solids, and flavor of kernels of sweet (cv. ‘Golden Cross Bantam 70’), super sweet (cv. ‘Cocktail E-51’), and waxy corns (cv. ‘Chalok 2’) were observed at different ripening stages. Sucrose was a major sugar in the sweet and super sweet corns and the content increased from 15 to 21 and 27 days after silking (DAS), respectively and then decreased. Glucose and fructose contents of sweet and super sweet corns tended to decrease with kernel maturity. Total sugar content of the sweet corn analyzed by the anthrone method increased rapidly from 15 to 21 DAS, while that of the super sweet and the waxy corns increased slowly up to 24 and 26 DAS, respectively and decreased thereafter. The content of soluble solids in sweet corn was much higher than that of super sweet corn. Starch content of the sweet corn increased slowly from 15 to 33 DAS, while that of the super sweet corn increased a little rapidly from 15 to 21 DAS and then leveled off to 33 DAS. Starch content of the waxy corn increased continuously from 21 to 38 DAS. There was a positive correlation between the sum of individual sugars (sucrose, glucose, and fructose) and soluble solids in both sweet and super sweet corns, while the content of soluble solids was not related to the sum of individual sugars or total sugars. The flavor rate of sweet and super sweet corns maintained high between 21 and 27 DAS and that of waxy corn decreased from 24 to 33 DAS. The optimum harvest time for sweet, super sweet, and waxy corns was thought to be 21 to 24 DAS considering sugar and starch contents, flavor, and marketing.

• 영재교육연구
• /
• 제16권1호
• /
• pp.1-19
• /
• 2006

본 연구는 최근 일본에서 실시된 고등학생을 대상으로 한 수학 과학 중심의 슈퍼 사이언스 하이 스쿨(Super Science High School ; 이하 SSH)의 실천 결과를 분석하여 그 성과와 문제점을 살펴본다. 이를 위해 본 연구는 2002년부터 2004년까지 SSH를 실시한 전국의 31개의 학교가 발행한 실천보고서를 대상으로 하여 항목별로 나누어 분석하였으며, 이를 통하여 우리나라 과학 고등학교의 수학 과학 교육에 대한 시사점을 얻고자 한다.

• Journal of Mechanical Science and Technology
• /
• 제19권12호
• /
• pp.2303-2311
• /
• 2005

Super low temperature has many applications nowadays, from the chemical processing, automotives manufacturing, plastic recycling, etc. Considering of its wide application in the present and the future, study of the super-low temperature refrigeration system should be actively carried out. Super low state temperature can be achieved by using multi-stage refrigeration system. This paper present the development and testing of cascade refrigerator system for achieving super-low temperature. On this experiment, two different types of HCFCs refrigerants are utilized, R-22 and R-23 were applied for the high stage and the low-pressure stage respectively. The lowest temperature in the low-pressure evaporator that can be achieved by this cascade refrigeration system is down to $-85^{\circ}C$. This experiment is aimed to study the effect of inlet pressure of the low-pressure stage evaporator and low-pressure stage compressors inlet pressure characteristics to the overall temperature characteristics of cascade refrigeration system.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 춘계학술대회
• /
• pp.129-132
• /
• 2022

초해상화란 저화질의 이미지를 고화질의 이미지로 변환하는 과정이다. 본 연구에서는 ESPCN 을 이용하여 연구를 진행하였다. 초해상화 심층 신경망에서 각 노드를 거칠 때 가중치를 결정하는 활성화 함수에 따라 같은 입력 데이터를 받더라도 다른 품질의 이미지가 출력될 수 있다. 따라서 활성화 함수 ReLU, ELU, Swish를 적용시켜 같은 입력 이미지에 대한 출력 이미지의 품질을 비교하여 초해상화에 가장 적합한 활성화 함수를 찾는 것이 이 연구의 목적이다. 초해상화를 위한 Dataset은 BSDS500 Dataset을 사용하였으며, 전처리 과정에서 이미지를 정사각형으로 자른 뒤 저화질화 하였다. 저화질화된 이미지는 모델의 입력 이미지에 사용되었고, 원본 이미지는 이후 출력 이미지와 비교하여 평가하는데 사용되었다. 학습 결과 머신 러닝에 주로 쓰이는 ReLU보다는 그 단점이 개선된 ELU, swish가 훈련 시간은 오래 걸렸지만 좋은 성능을 보였다.