• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2017.04a
• /
• pp.162-162
• /
• 2017

구글에서 공개한 Tensorflow를 이용한 여러 학문 분야의 연구가 활발하다. 농업 시설환경을 대상으로 한 빅데이터의 축적이 증가함과 아울러 실효적인 정보 획득을 위한 각종 데이터 분석 및 마이닝 기법에 대한 연구 또한 활발한 상황이다. 한편, 타 분야의 성공적인 심층학습기법 응용사례에 비하여 농업 분야에서의 응용은 초기 성장 단계라 할 수 있다. 이는 농업 현장에서 취득한 정보의 난해성 및 완성도 높은 생육/환경 모델링 정보의 부재로 실효적인 전과정 처리 기술 도출에 소요되는 시간, 비용, 연구 환경이 상대적으로 부족하기 때문일 것이다. 특히, 센서 기반 데이터 취득 기술 증가에 따라 비약적으로 방대해진 수집 데이터를 시간 복잡도가 높은 심층 학습 모델링 연산에 기계적으로 단순 적용할 경우 시간 효율적인 측면에서 성공적인 결과 도출에 애로가 있을 것이다. 매우 높은 시간 복잡도를 해결하기 위하여 제시된 하드웨어 가속 기능의 경우 일부 개발환경에 국한이 되어 있다. 일례로, 구글의 Tensorflow는 오픈소스 기반 병렬 클러스터링 기술인 MPICH를 지원하는 알고리즘을 공개하지 않고 있다. 따라서, 본 연구에서는 심층학습 기법 연구에 있어서, 예상 가능한 다양한 자원을 활용하여 최대한 연산의 결과를 빨리 도출할 수 있는 하드웨어적인 접근 방법을 모색하였다. 호스트에서 수행하는 일방적인 학습 알고리즘과 달리 이기종간 심층 학습이 가능하기 위해선 우선, NFS(Network File System)를 이용하여 데이터 계층이 상호 연결이 되어야 한다. 이를 위해서 고속 네트워크를 기반으로 한 NFS의 이용이 필수적이다. 둘째로 제한된 자원의 한계를 극복하기 위한 메모 공유 라이브러리가 필요하다. 셋째로 이기종간 프로세서에 최적화된 병렬 처리용 컴파일러를 이용해야 한다. 가장 중요한 부분은 이기종간의 처리 능력에 따른 작업을 고르게 분배할 수 있는 작업 스케쥴링이 수행되어야 하며, 이는 처리하고자 하는 데이터의 형태에 따라 매우 가변적이므로 해당 데이터 도메인에 대한 엄밀한 사전 벤치마킹이 수행되어야 한다. 이러한 요구조건을 대부분 충족하는 Open-CL ver1.2(https://www.khronos.org/opencl/)를 이용하였다. 최신의 Open-CL 버전은 2.2이나 본 연구를 위하여 준비한 4가지 이기종 시스템에서 모두 공통적으로 지원하는 버전은 1.2이다. 실험적으로 선정된 4가지 이기종 시스템은 1) Windows 10 Pro, 2) Linux-Ubuntu 16.04.4 LTS-x86_64, 3) MAC OS X 10.11 4) Linux-Ubuntu 16.04.4 LTS-ARM Cortext-A15 이다. 비교 분석을 위하여 NVIDIA 사에서 제공하는 Pascal Titan X 2식을 SLI로 구성한 시스템을 준비하였다. 개별 시스템에서 별도로 컴파일 된 바이너리의 이름을 통일하고, 개별 시스템의 코어수를 동일하게 균등 배분하여 100 Hz의 데이터로 입력이 되는 온도 정보와 조도 정보를 입력으로 하고 이를 습도정보에 Linear Gradient Descent Optimizer를 이용하여 Epoch 10,000회의 학습을 수행하였다. 4종의 이기종에서 총 32개의 코어를 이용한 학습에서 17초 내외로 연산 수행을 마쳤으나, 비교 시스템에서는 11초 내외로 연산을 마치는 결과가 나왔다. 기보유 하드웨어의 적절한 활용이 가능한 심층학습 기법에 대한 연구를 지속할 것이다

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.3
• /
• pp.271-278
• /
• 2014

In spite of that stop line and crosswalk should be aware of the most basic objects in transportation system, its features extracted are very limited. In addition to image-based recognition technology, laser and RF, GPS/INS recognition technology, it is difficult to recognize. For this reason, the limited research in this area has been done. In this paper, the algorithm to recognize the stop line and crosswalk is designed and implemented using image-based recognition technology with the images input through a vision sensor. This algorithm consists of three functions.; One is to select the area, in advance, needed for feature extraction in order to speed up the data processing, 'Region of Interest', another is to process the images only that white color is detected more than a certain proportion in order to remove the unnecessary operation, 'Color Pattern Inspection', the other is 'Feature Extraction and Recognition', which is to extract the edge features and compare this to the previously-modeled one to identify the stop line and crosswalk. For this, especially by using case based feature comparison algorithm, it can identify either both stop line and crosswalk exist or just one exists. Also the proposed algorithm is to develop existing researches by comparing and analysing effect of in-vehicle camera installation and changes in recognition rate of distance estimation and various constraints such as backlight and shadow.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.12
• /
• pp.271-279
• /
• 2019

In extreme environment regions, unmanned rovers equipped with various sensors and devices are being developed for long-term exploration on behalf of humans. On the other hand, due to the harsh weather conditions and rough terrain, the rover camera has limited visible distance and field of view. Therefore, the rover cameras should be located for safe navigation and efficient terrain mapping. In this regard, to minimize the cost and time to manufacture the camera system on a rover, the simulation method using the rover design is presented to optimize the camera locations on the rover efficiently. In the simulation, a simulated terrain was taken from cameras with different locations and angles. The visible distance and overlapped extent of camera images, and terrain data accuracy calculated from the simulation were compared to determine the optimal locations of the rover's cameras. The simulated results will be used to manufacture a rover and camera system. In addition, self and system calibrations will be conducted to calculate the accurate position of the camera system on the rover.

• Korean Journal of Remote Sensing
• /
• v.30 no.1
• /
• pp.47-59
• /
• 2014

Volcanic ash (VA) can be estimated by remote sensing sensors through their spectral signatures determined by the inherent optical property (IOP) including complex refractive index and the scattering properties. Until now, a very limited range of VA refractive indices has been reported and the VA from each volcanic eruption has a different composition. To improve the robustness of VA remote sensing, there is a need to understanding of VA - radiation interactions. In this study, we calculated extinction coefficient, scattering phase function, asymmetry factor, and single scattering albedo which show different values between andesite and pumice. Then, IOPs were used to analyze the relationship between theoretical remote sensed radiation calculated by radiative transfer model under various aerosol optical thickness (${\tau}$) and sun-sensor geometries and characteristics of VA. It was found that the mean rate of change of radiance at top of atmosphere versus ${\tau}$ is six times larger than in radiance values at 0.55 ${\mu}m$. At the surface, positive correlation dominates when ${\tau}$ <1, but negative correlation dominates when ${\tau}$ >1. However, radiance differences between andesite and pumice at 11 ${\mu}m$ are very small. These differences between two VA types are expressed as the polynomial regression functions and that increase as VA optical thickness increases. Finally, these results would allow VA to be better characterized by remote sensing sensors.

• Journal of Korean Society of Transportation
• /
• v.35 no.2
• /
• pp.116-128
• /
• 2017

Traffic accidents can be defined as a physical collision event of vehicles occurred instantaneously when drivers do not perceive the surrounding vehicles and roadway environments properly. Therefore, detecting the high potential events that cause traffic accidents with monitoring the interactions among the surroundings continuously by driver is the prerequisite for prevention the traffic accidents. For the analysis, basic data were collected to analyze interactions using a test vehicle which is equipped the GPS(Global Positioning System)-IMU(Inertial Measurement Unit), camera, radar and RiDAR. From the collected data, highway geometric information and the surrounding traffic situation were analyzed and then safety evaluation algorithm for driving vehicle was developed. In order to detect a dangerous event of interaction with surrounding vehicles, locations and speed data of surrounding vehicles acquired from the radar sensor were used. Using the collected data, the tangent and curve section were divided and the driving safety evaluation algorithm which is considered the highway geometric characteristic were developed. This study also proposed an algorithm that can assess the possibility of collision against surrounding vehicles considering the characteristics of geometric road structure. The methodology proposed in this study is expected to be utilized in the fields of autonomous vehicles in the future since this methodology can assess the driving safety using collectible data from vehicle's sensors.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.8 no.6
• /
• pp.907-915
• /
• 2013

Tire Pressure Monitoring System (TPMS) is an auxiliary safety system for recognizing the condition of tires based on the pressure and temperature data transmitted from the sensor unit installed on a tire of the vehicle. Using TPMS, a driver can frequently check the state of tires and it aids to maintain the optimum running condition of the vehicle. Since TPMS must utilize the wireless communication technique to transmit data from a sensor unit to a signal processing unit installed in the vehicle, it suffers from interference signals caused by various external electrical or electronic devices. In order to suppress high-power interference signals, we employ beamforming techniques based on the uniform linear antenna array. As the number of the antennas is increased, the performance of the interference suppression is improved. However, there is the limit of the number of antennas, installed in the center of a vehicle, because of its size. In this paper, we compare and analyze the performance of the beamformer, when reducing the interelement spacing of antennas, to increase the number of the receiving antennas. For the performance analysis of the beamformers, we consider the switching beamformer and minimum-variance distortionless-response (MVDR) beamformer for TPMS, recently proposed.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2010.06a
• /
• pp.18-18
• /
• 2010

전기적 에너지를 기계적 에너지로 변환하고 또한, 기계적 에너지를 전기적 에너지로 변환할 수 있는 압전 세라믹스는 압전 변압기 (piezoelectric transformer), 초음파 모터, 센서 등과 같은 응용분야에 광범위하게 사용되고 있다. 특히, 전원장치에 있어서 현재 주요 전자제품에 사용되고 있는 권선형 변압기와 같은 전자 변환기의 대체품으로 압전 세라믹스 소재의 특성을 이용한 압전변압기의 개발과 응용연구는 국내외적으로 활발히 연구되어왔다. 압전변압기는 권선형 변압기와 비교 하였을 때 누설자속이 없어 노이즈 발생이 없고, 공진주파수만을 이용하므로 출력의 파형이 정현파에 가까워 고조파 잡음이 없으며, 불연성의 특징을 가지고 있다. 추가적으로 압전 변압기는 소형화, 슬림화, 경량화가 가능하며 90%이상의 높은 효율을 얻을 수 있다. 또한, 단판형 압전변압기의 출력한계를 개선하기 위해 높은 승압비와 고출력을 갖는 적층타입의 압전변압기가 제안되었다. 압전변압기용 조성 세라믹스는 높은 에너지 변환효율을 위해서 전기기계결합계수 ($k_p$)가 커야 하며, 발열에 의한 온도 상승을 억제하기 위하여 기계적 품질계수(Qm)가 큰 것이 바람직하다. 또한, 높은 전류를 발생하기 위해서는 유전상수가 커 압전변압기의 출력측 정전용량을 크게 하여야한다. 이러한 압전변압기의 제작 조건을 위해 우수한 압전 및 유전특성을 갖는 PZT계 세라믹스가 주로 사용 되어져 왔다. 그러나, PZT계 세라믹스의 우수한 압전 및 유전특성에도 불구하고 $1000^{\circ}C$에서 급격히 휘발하는 PbO의 성질 때문에 환경적으로나 인체의 건강문제로 인해 전세계적으로 그 사용량을 제한하고 있다. 또한 적층 압전변압기의 구조적 특성상 내부전극과 함께 소결하여야 하는데, 이때 소결온도가 높으면 값비싼 Pd합량이 높은 전극을 사용하여야 한다. Pd함량이 10%미만인 Ag/Pd 전극을 사용하기 위해서는 $950^{\circ}C$ 이하에서 저온소결이 가능한 세라믹스 제조가 필수적이라 할 수 있다. 소결온도를 낮추는 방법으로는 다른 물질들을 치환하여 소결온도를 낮추는 방법과 미세분말을 만들어 그레인사이즈를 미세화 하는 방법들이 있다. 많은 미세 분말 제조 방법 중에서 Attrition mill은 일반적인 ball mill에 비해 분말의 입도를 미세하게 할 수 있어 증가된 분말의 비표면적에 의하여 반응을 촉진시킴으로써 저온소결이 가능한 세라믹스를 만들 수 있다. 따라서 본 연구에서는 소결온도가 낮으면서도 유전 및 압전특성이 우수한 조성을 사용하여 적층 압전변압기를 제작하여 전기적 특성을 조사하였다.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.12
• /
• pp.210-216
• /
• 2014

In this paper, it was proposed to develop front and rear image monitoring system for vehicle that help a driver to cope with urgent situation about a dangerous element. When parking a vehicle, the risk factors to be formed by the dead zone can be resolved by using anterior and posterior cameras of the vehicle. In embedded system environment, a SoC(System on Chip) and two high-resolution CMOS (Complementary metal-oxide-semiconductor) image sensors were used to transfer two high-resolution image data through he TCP/ IP-based network. To transfer image data through he TCP/ IP-based network, the images received by two cameras were compressed by using H.264 and they were transmitted with wireless method(Wi-Fi) by using real-time transport protocol (Real-time Transport Protocol). Transmission loss, transmission delay and transmission limit were solved in wireless (Wi-Fi) environment and the bit-rate of two image data compressed by H.264 was adjusted. And the system for the optimal transmission in wireless (Wi-Fi) environment was materialized and experimented.

• Journal of the Society of Cosmetic Scientists of Korea
• /
• v.30 no.2
• /
• pp.173-180
• /
• 2004

Nano sized ZnO particle as 20-30nm applies for material, pigments, rubber additives, gas sensors, varistors, fluorescent substance as well as new material such as photo-catalyst, sensitizer, fluorescent material. ZnO with a particle size in the range 20-30nm has provided to be an excellent UV blocking material in the cosmetics industry, which can be used in sunscreen product to enhance the sun protection factor and natural makeup effect. But pure ZnO particles application limits for getting worse wearing feeling. We make high-functional inorganic-composite that coated with nano-ZnO on the plate-type particle such as sericite, boron nitride and bismuthoxychloride. In this experiment, we synthesized composite powder using hydrothermal precipitation method. The starting material was ZnCl$_2$ Precipitation materials were used hexamethylenetetramine(HMT) and urea. We make an experiment with changing as synthesis factors that are concentrations of starting material, precipitation materials, nuclear formation material, reaction time, and reaction temperature. We analyzed composite powder's shape, crystallization and UV-blocking ability with FE-SEM, XRD, FT-IR, TGA-DTA, In vitro SPF test. The user test was conducted by product's formulator. In the results of this study, nanometer sized ZnD was coated regardless of the type of plate-powder at fixed condition range. When the coated plate-powders were applied in pressed powder product, the glaze of powder itself decreased, but natural make-up effect, spreadability, and adhesionability were increased.

• Korean Journal of Remote Sensing
• /
• v.14 no.3
• /
• pp.277-284
• /
• 1998

To establish a monthly data collection planning for the Ocean Scanning Multispectral Imager (OSMI), we have examined the global patterns of three impacting factors: pigment concentration, cloud cover, and sun glint. Other than satellite mission constraints (e.g., duty cycle), these three factors are considered critical for the OSMI data collection. The Nimbus-7 Coastal Zone Color Scanner (CZCS) monthly mean products and the International Satellite Cloud Climatology Project (ISCCP) monthly mean products (C2) were used for the analysis of pigment concentration and cloud cover distributions, respectively. And the monthly-simulated patterns of sun glint were produced by performing the OSMI orbit prediction and the calculation of sun glint radiances at the top-of-atmosphere (TOA). Using monthly statistics (mean and/or standard deviation) of each factor in the above for a given 10$^{\circ}$ latitude by 10$^{\circ}$ longitude grid, we generated the priority map for each month. The priority maps of three factors for each month were subsequently superimposed to visualize the impact of three factors in all. The initial results illustrated that a large part of oceans in the summer hemisphere was classified into the low priority regions because of seasonal changes of clouds and sun illumination. Sensitivity tests for different sets of classifications were performed and demonstrated the seasonal effects of clouds and sun glint to be robust.