• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.2
• /
• pp.10-23
• /
• 2009

The fundamental matrix and key frame selection are one of the most important techniques to recover full 3D reconstruction of objects from turntable sequences. This paper proposes a new algorithm that estimates a robust fundamental matrix for camera calibration from uncalibrated images taken under turn-table motion. Single axis turntable motion can be described in terms of its fixed entities. This provides new algorithms for computing the fundamental matrix. From the projective properties of the conics and fundamental matrix the Euclidean 3D coordinates of a point are obtained from geometric locus of the image points trajectories. Experimental results on real and virtual image sequences demonstrate good object reconstructions.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.627-629
• /
• 2023

생체 신호를 분석하여 사용자의 건강과 정신 상태를 예측하고, 관련 질병에 관해 예방하는 연구가 늘어나고 있다. 생체 신호 중 심박은 사람의 육체, 정신적인 상태를 반영하는 대표적인 신호이지만 기존의 접촉 패드를 통한 ECG나 광학 센서를 통한 PPG로 심박을 예측할 때는 구속적인 환경이 필요하여 일상적인 상황 속에 적용하기 어려웠다. 이러한 단점을 해결하고자 본 논문은 UBFC-RPPG 데이터셋의 동영상 프레임을 RGB 채널마다 다른 가중치를 적용하는 전처리를 하여 학습 데이터의 크기를 줄이면서 정확도를 높이고, 3D-CNN을 활용한 딥러닝으로 순간적인 영상에서도 PPG 신호를 예측할 수 있도록 1초 전처리 영상을 학습한 후, 신호를 예측하는 것을 목표로 한다. 이렇게 비접촉식으로 취득된 신호는 더 다양한 환경에서의 감정분류, 우울증 진단, 질병 감지 등 다양한 분야에 활용될 수 있다.

• Journal of Korean Society for Geospatial Information Science
• /
• v.16 no.1
• /
• pp.67-73
• /
• 2008

The most common method of pipeline inspection is to use a remote-controlled-machine equipped with a CCTV, which, however, has many limitations to accurately inspect pipeline condition. In case of a typical CCTV, since the camera looks at the end point of the pipe, the locations of the defects and distance-readings are often different. In addition, the quality and accuracy of the inspection is highly dependent on the operator's skill and experience. In this research a new system is developed by use of the image processing techniques and the lens combination. The image acquisition system is developed that acquires the front and the side view of the pipe simultaneously. Side view unwrapping and stitching technology using image process techniques are developed which delivers high resolution image data.

• Korean Journal of Remote Sensing
• /
• v.32 no.5
• /
• pp.511-527
• /
• 2016

We propose a rotating stereo frame camera system to acquire indoor images with a low cost. For experiments, we selected a test site and acquired images using the proposed system and control points using a total station. Using these data, we generated various indoor 3D models using commercial photogrammetric software, PhotoScan. We then performed qualitative and quantitative analysis of the generated indoor 3D models to investigate the possibility of the indoor modeling using the proposed system. From the results, it is confirmed that the generated indoor models using the proposed system can be applicable to the services not inquiring high accuracy.

• Journal of IKEEE
• /
• v.18 no.4
• /
• pp.612-619
• /
• 2014

This paper proposes a real-time multi-camera video surveillance system with automatic resolution control using motion detection. In ordinary times, it acquires 4 channels of QVGA images, and it merges them into single VGA image and transmit it. When motion is detected, it automatically increases the resolution of motion-occurring channel to VGA and decreases those of 3 other channels to QQVGA, and then these images are overlaid and transmitted. Thus, it can magnifies and watches the motion-occurring channel while maintaining transmission bandwidth and monitoring all other channels. When it is synthesized with 0.18 um technology, the maximum operating frequency is 110 MHz, which can theoretically support 4 HD cameras.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10b
• /
• pp.544-546
• /
• 2003

홍채인식 기술에서 홍채영역 추출 방법은 필요한 데이터를 추출, 변환하는 과정에서 발생하는 홍채정보의 손실을 최소화하는데 그 목적이 있다. 그러나 기존의 홍채인식에서 사용되는 홍채영역 추출 방법은 초기 눈 영상 획득 시 얻은 영상의 눈 기울기가 일정하다는 가정에 기초하거나, 하나의 홍채 데이터를 기준으로 삼고, 다른 홍채데이터를 비교하여 기울기를 보정하기 때문에 동일인에게서 취득한 영상이라 하더라도 영상간의 기울기가 크다면 두 홍채 데이터의 공통된 특징 역시 상대적으로 적거나 차이가 날 수밖에 없다. 결국, 이는 실제 필드에서 사용할 시스템의 인식률에 영향을 줄 수밖에 없다. 본 논문에서는 이 문제를 보완하기 위한 방안으로, 눈 영상 내에서 내안각을 찾아내고 이를 기준으로 하여, 눈 영상의 특정 부분에서 일정한 홍채영역을 추출하는 방법을 제안한다. 본 논문에서는 우리가 제안하는 방법을 사용하여 실제 취득한 모든 눈 영상에 대한 실험을 한 결과, 비교적 일정한 홍채영역을 추출할 수 있음을 입증하였다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.158-158
• /
• 2020

국내외 기상이변으로 인한 하천허용유량의 초과가 많이 일어나고 있다. 이때 사람이 직접 도섭으로 하천의 정보 대한 정보를 취득하기 어렵기 때문에 간접적인 방법으로의 정보취득이 중요시 되어야 한다. 그에 따른 방법으로 주입식의 표면부자를 활용한 방법이나, PIV 기법을 활용한 표면영상유속계측법을 이용한 유속측정방법 등이 있다. 하지만 이런 표면영상유속계측법은 영상의 변위가 커지게 되어 시간간격 설정에 대한 어려움 등이 있고, 표면부자법의 경우 표면 유속을 대변 할 수 있을지에 대한 분석연구가 필요하다. 따라서 이러한 표면 영상 유속계측법과 표면부자의 수표면 유속 측정 능력 등을 분석하여 GPS 전자부자의 활용성을 확인해 보고자 하였다. Lab-Scale 수로에서 GPS전자부자를 입자로 활용하여 주입 후 흘러가는 부분을 영상촬영하고, LS-PIV 기법을 활용하여 분석한 후 검증이 된 전자유속계 (FlowTracker)의 데이터를 참값으로 비교 분석하였다. LS-PIV 기법을 통해서 GPS 전자부자의 표면 유속 측정 능력을 잘 검증해낸다면, 유속측정에 용이하고 접근성 및 안정성까지 확보할 수 있다. 더 나아가 홍수 시 사용뿐만 아니라, ADCP나 ADV로 측정이 불가능한 저수심·저유속의 중소하천에서도 적용할 수 있다. GPS전자부자의 지속적인 연구는 향후 하천 측정 시스템에 많은 도움을 줄 것으로 사료된다.

• Korean Journal of Remote Sensing
• /
• v.36 no.5_3
• /
• pp.1013-1025
• /
• 2020

Unmanned Aerial Vehicle (UAV) platform is being widely used in disaster monitoring and smart city, having the advantage of being able to quickly acquire images in small areas at a low cost. Ground Control Points (GCPs) for positioning UAV images are essential to acquire cm-level accuracy when producing UAV-based orthoimages and Digital Surface Model (DSM). However, the on-site acquisition of GCPs takes considerable manpower and time. This research aims to provide an efficient and accurate way to replace the on-site GNSS surveying with three different sources of geospatial data. The three geospatial data used in this study is as follows; 1) 25 cm aerial orthoimages, and Digital Elevation Model (DEM) based on 1:1000 digital topographic map, 2) point cloud data acquired by Mobile Mapping System (MMS), and 3) hybrid point cloud data created by merging MMS data with UAV data. For each dataset a three-dimensional positional accuracy analysis of UAV-based orthoimage and DSM was performed by comparing differences in three-dimensional coordinates of independent check point obtained with those of the RTK-GNSS survey. The result shows the third case, in which MMS data and UAV data combined, to be the most accurate, showing an RMSE accuracy of 8.9 cm in horizontal and 24.5 cm in vertical, respectively. In addition, it has been shown that the distribution of geospatial GCPs has more sensitive on the vertical accuracy than on horizontal accuracy.

• The KIPS Transactions:PartB
• /
• v.9B no.4
• /
• pp.453-458
• /
• 2002

It is not easy to analyze dynamic sports motion, but it can be done easily by using an interactive playback of the recorded motion. Everybody is able to capture a golf swing owing to the popularization of PC and digital camcorder in these days. In this study, the system that can provide golf swing analysis is selected and implemented. Furthermore, editing method that can represent motion is applied and moving object tracking algorithm that can make this profess easy is applied to provide easy golf swing analysis.

• International Journal of Highway Engineering
• /
• v.7 no.4 s.26
• /
• pp.163-171
• /
• 2005

Current roads were designed and constructed based on the design criteria and thus those were overly simplified drivers' needs. The road criteria do not suggest the desirable range of the design values but suggest the minimum requirements for the road design. Therefore, a completed road design based on the design criteria does not always guarantee the best design in terms of safety and it sometimes violates drivers' expectation. Therefore, the ROSSAV(ROad Safety Survey and Analysis Vehicle) is being developed by the KICT to evaluate road safety and increase driving safety. In this paper, the image capture system was described in detail. The image capture system is consisted of two front view cameras, two side down-looking cameras and a synchronization device. Two front view cameras were used to take a picture of road and road facilities at the driver's viewpoint. Also, two side down-looking cameras were used to capture road surface image to extract lane markings. A synchronization device were used to generate image capturing signal at the fixed distance spacing huck as every 10m. The front view images could be used to calculate and measure highway geometry such as shoulder width because every image is saved with it's locational information. And also the side down looking images could be used to extract median lane mark which representing road alignement efficiently.