• International Journal of Aeronautical and Space Sciences
• /
• v.17 no.2
• /
• pp.232-239
• /
• 2016

Recently, in order to increase the efficiency and mission success rate of UAVs (Unmanned Aerial Vehicles), the necessity for formation flights is increased. In general, GPS (Global Positioning System) is used to obtain the relative position of leader with respect to follower in formation flight. However, it can't be utilized in environment where GPS jamming may occur or communication is impossible. Therefore, in this study, monocular vision is used for measuring relative position. General PC-based vision processing systems has larger size than embedded systems and is hard to install on small vehicles. Thus FPGA-based processing board is used to make our system small and compact. The processing system is divided into two blocks, PL(Programmable Logic) and PS(Processing system). PL is consisted of many parallel logic arrays and it can handle large amount of data fast, and it is designed in hardware-wise. PS is consisted of conventional processing unit like ARM processor in hardware-wise and sequential processing algorithm is installed on it. Consequentially HW/SW co-designed FPGA system is used for processing input images and measuring a relative 3D position of the leader, and this system showed RMSE accuracy of 0.42 cm ~ 0.51 cm.

• The Journal of Korea Robotics Society
• /
• v.6 no.3
• /
• pp.197-209
• /
• 2011

This paper presents a localization system using ceiling images in a large indoor environment. For a system with low cost and complexity, we propose a single camera based system that utilizes ceiling images acquired from a camera installed to point upwards. For reliable operation, we propose a method using hybrid features which include natural landmarks in a natural scene and artificial landmarks observable in an infrared ray domain. Compared with previous works utilizing only infrared based features, our method reduces the required number of artificial features as we exploit both natural and artificial features. In addition, compared with previous works using only natural scene, our method has an advantage in the convergence speed and robustness as an observation of an artificial feature provides a crucial clue for robot pose estimation. In an experiment with challenging situations in a real environment, our method was performed impressively in terms of the robustness and accuracy. To our knowledge, our method is the first ceiling vision based localization method using features from both visible and infrared rays domains. Our system can be easily utilized with a variety of service robot applications in a large indoor environment.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.05a
• /
• pp.91-93
• /
• 2016

From the 2-D image extracting three-dimensional information as the latter is in the bilateral sibeop using two camera method and when using a monocular camera as a very important step generally as "stereo vision". There in today's CCTV and automatic object tracking system used in many medium much to know the site conditions or work developed more clearly by using a stereo camera that mimics the eyes of humans to maximize the efficiency of avoidance / control start and multiple jobs can do. Object tracking system of the existing 2D image will have but can not recognize the distance to the transition could not be recognized by the observer display using a parallax of a stereo image, and the object can be more effectively controlled.

• Journal of Digital Contents Society
• /
• v.11 no.4
• /
• pp.597-606
• /
• 2010

This paper describes a real-time camera tracking framework specifically designed to track a monocular camera in an AR workspace. Typically, the Kalman filter is often employed for the camera tracking. In general, however, tracking performances of conventional methods are seriously affected by unpredictable situations such as ambiguity in feature detection, occlusion of features and rapid camera shake. In this paper, a recursive Bayesian sampling framework which is also known as the particle filter is adopted for the camera pose estimation. In our system, the camera state is estimated on the basis of the Gaussian distribution without employing additional uncertainty model and sample weight computation. In addition, the camera state is directly computed based on new sample particles which are distributed according to the true posterior of system state. In order to verify the proposed system, we conduct several experiments for unstable situations in the desktop AR environments.

• Annals of Clinical Neurophysiology
• /
• v.2 no.1
• /
• pp.13-20
• /
• 2000

Background and Objective : Visual evoked potentials(VEPs) is considered to be a reliable diagnostic procedure for examining patients with anterior visual pathways. Some abnormalities in the recordings on monocular stimulation have been said to indicate retrochiasmal lesion, but less consistent results have been reported. This study is to evaluate the positive predictability of VEP for the detection of retrochiasmal lesion. Methods : We reviewed VEPs that could be interpreted as indicative of a retrochiasmal lesions, based on amplitude or latency asymmetry recorded on the left(O1) and right(O2) occipital regions. Bilateral absent VEPs on both recording(O1 and O2) without evidence of prechiasmal lesion were included. During 5 years, we identified 31 patients who met the above criteria and who had undergone magnetic resonance imaging(MRI) of brain(one patient underwent computerized tomography). Twenty three patients underwent pattern reversal VEPs and others underwent flash goggle VEPs. Results : Brain imagings were abnormal in 29 and were normal in 2. Of the 29 abnormal scans, lesions in posterior visual pathway were detected in 21 scans(predictive value=68%). The predictive value was not significantly different between flash goggle VEP(75%) and pattern reversal VEP(68%). The predictive value was higher in patient with visual field defect(100%) than those without visual field defect(25%). The pathologic nature of lesion also showed close relations to the predictive value. VEPs is usually paradoxically lateralized(78%), but not in all patients. Conclusion : VEPs abnormalities suggesting retrochiasmal lesion were usually corresponded with brain MRI findings. Diagnostic reliability could be increased when considering the visual field defect and nature of lesion. Therefore, the authors suggest that VEPs studies could be useful in evaluating the patients with the retrochismal lesion.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.7
• /
• pp.2579-2585
• /
• 2010

This paper proposes a dynamic human pose tracking method using motion-based search strategy from an image sequence obtained from a monocular camera. The proposed method compares the image features between 3D human model projections and real input images. The method repeats the process until predefined criteria and then estimates 3D human pose that generates the best match. When searching for the best matching configuration with respect to the input image, the search region is determined from the estimated 2D image motion and then search is performed randomly for the body configuration conducted within that search region. As the 2D image motion is highly constrained, this significantly reduces the dimensionality of the feasible space. This strategy have two advantages: the motion estimation leads to an efficient allocation of the search space, and the pose estimation method is adaptive to various kinds of motion.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.10a
• /
• pp.357-359
• /
• 2010

In monocular images that are used to determine the depth of the vanishing point, the buildings, roads and buildings, such as outdoor video or hallway with room inside for the interior structure, such as the vanishing point in the video is a very strong depth cue. Depth map using the vanishing point in the three-dimensional space, the two-dimensional imaging is used to restore the structure. But if there is a vanishing point vanishing point in the video also depends on the location of the relative depth of different ways to express that need. In this paper we present images of a vanishing point with respect to the improved depth-map was created. Proposed an area where the loss of seven points and areas defined as areas along the proposed direction of different depth.

• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.70-78
• /
• 2021

Estimating 3D information from a single image is one of the essential problems in numerous applications. Since a 2D image inherently might originate from an infinite number of different 3D scenes, thus 3D reconstruction from a single image is notoriously challenging. This challenge has been overcame by the advent of recent deep convolutional neural networks (CNNs), by modeling the mapping function between 2D image and 3D information. However, to train such deep CNNs, a massive training data is demanded, but such data is difficult to achieve or even impossible to build. Recent trends thus aim to present deep learning techniques that can be trained in a weakly-supervised manner, with a meta-data without relying on the ground-truth depth data. In this article, we introduce recent developments of weakly-supervised deep learning technique, especially categorized as scene 3D reconstruction and object 3D reconstruction, and discuss limitations and further directions.

• Journal of Broadcast Engineering
• /
• v.27 no.3
• /
• pp.341-350
• /
• 2022

In this paper, we propose a deep learning based on-device augmented reality (AR) system in which multiple input images are used to implement the correct occlusion in a real environment. The proposed system is composed of three technical steps; camera pose estimation, depth estimation, and object augmentation. Each step employs various mobile frameworks to optimize the processing on the on-device environment. Firstly, in the camera pose estimation stage, the massive computation involved in feature extraction is parallelized using OpenCL which is the GPU parallelization framework. Next, in depth estimation, monocular and multiple image-based depth image inference is accelerated using the mobile deep learning framework, i.e. TensorFlow Lite. Finally, object augmentation and occlusion handling are performed on the OpenGL ES mobile graphics framework. The proposed augmented reality system is implemented as an application in the Android environment. We evaluate the performance of the proposed system in terms of augmentation accuracy and the processing time in the mobile as well as PC environments.

• Korean Journal of Remote Sensing
• /
• v.39 no.1
• /
• pp.1-21
• /
• 2023

In this technically advanced era, we are surrounded by smartphones, computers, and cameras, which help us to store visual information in 2D image planes. However, such images lack 3D spatial information about the scene, which is very useful for scientists, surveyors, engineers, and even robots. To tackle such problems, depth maps are generated for respective image planes. Depth maps or depth images are single image metric which carries the information in three-dimensional axes, i.e., xyz coordinates, where z is the object's distance from camera axes. For many applications, including augmented reality, object tracking, segmentation, scene reconstruction, distance measurement, autonomous navigation, and autonomous driving, depth estimation is a fundamental task. Much of the work has been done to calculate depth maps. We reviewed the status of depth map estimation using different techniques from several papers, study areas, and models applied over the last 20 years. We surveyed different depth-mapping techniques based on traditional ways and newly developed deep-learning methods. The primary purpose of this study is to present a detailed review of the state-of-the-art traditional depth mapping techniques and recent deep learning methodologies. This study encompasses the critical points of each method from different perspectives, like datasets, procedures performed, types of algorithms, loss functions, and well-known evaluation metrics. Similarly, this paper also discusses the subdomains in each method, like supervised, unsupervised, and semi-supervised methods. We also elaborate on the challenges of different methods. At the conclusion of this study, we discussed new ideas for future research and studies in depth map research.