• Smart Media Journal
• /
• v.2 no.4
• /
• pp.27-33
• /
• 2013

Multi-view video system provides both realistic 3D feelings and free-view navigation. But it is hard to transmit too huge data, so we send only two or three view images and generate intermediate view image using depth information. In this paper, we propose high quality multi-view image generation method from stereoscopic images. Since the stereo matching method does not provide accurate disparity values for all the pixels, especially at the occlusion area, we propose an occlusion handling method using the background pixels at first. We also apply a joint bilateral filtering to enhance the disparity map at the object boundary since it can affect the quality of synthesized images significantly. Finally, we can generate virtual view images at intermediate view positions using confidence map to reduce bad pixel and hole's error. Experimental results show the proposed method performs better than the conventional method.

• Journal of Korean Society for Geospatial Information Science
• /
• v.14 no.2 s.36
• /
• pp.3-13
• /
• 2006

Extraction of man-made objects from high resolution satellite images has been studied by many researchers. In order to reconstruct accurate 3D building structures most of previous approaches assumed 3D information obtained by stereo analysis. For this, they need the process of sensor modeling, etc. We argue that a single image itself contains many clues of 3D information. The algorithm we propose projects virtual shadow on the image. When the shadow matches against the actual shadow, the height of a building can be determined. If the height of a building is determined, the algorithm draws vertical lines of sides of the building onto the building in the image. Then the roof boundary moves along vertical lines and the footprint of the building is extracted. The algorithm proposed can use the shadow cast onto the ground surface and onto facades of another building. This study compared the building heights determined by the algorithm proposed and those calculated by stereo analysis. As the results of verification, root mean square errors of building heights were about 1.5m.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.6A
• /
• pp.458-463
• /
• 2012

In this paper, we propose a direct depth map acquisition method for the convergence camera array as well as the parallel camera array. The conventional methods perform image rectification to reduce complexity and improve accuarcy. However, image rectification may lead to unwanted consequences for the convergence camera array. Thus, the proposed method excludes image rectification and directly extracts depth values using the epipolar constraint. In order to acquire a more accurate depth map, occlusion detection and handling processes are added. Reasonable depth values are assigned to the obtained occlusion region by the distance and color differences from neighboring pixels. Experimental results show that the proposed method has fewer limitations than the conventional methods and generates more accurate depth maps stably.

• International conference on construction engineering and project management
• /
• 2020.12a
• /
• pp.313-322
• /
• 2020

The recognition of the risk hazards is a vital step to effectively prevent accidents on a construction site. The advanced development in computer vision systems and the availability of the large visual database related to construction site made it possible to take quick action in the event of human error and disaster situations that may occur during management supervision. Therefore, it is necessary to analyze the risk factors that need to be managed at the construction site and review appropriate and effective technical methods for each risk factor. This research focuses on analyzing Occupational Safety and Health Agency (OSHA) related to risk zone identification rules that can be adopted by the image recognition technology and classify their risk factors depending on the effective technical method. Therefore, this research developed a pattern-oriented classification of OSHA rules that can employ a large scale of safety hazard recognition. This research uses joint reasoning of risk zone Identification and numeric input by utilizing a stereo camera integrated with an image detection algorithm such as (YOLOv3) and Pyramid Stereo Matching Network (PSMNet). The research result identifies risk zones and raises alarm if a target object enters this zone. It also determines numerical information of a target, which recognizes the length, spacing, and angle of the target. Applying image detection joint logic algorithms might leverage the speed and accuracy of hazard detection due to merging more than one factor to prevent accidents in the job site.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.2
• /
• pp.69-79
• /
• 1994

The conventional machine vision system which has uniform rectangular grid requires tremendous amount of computation for processing and analysing an image especially in 2-D image transfermations such as scaling, rotation and 3-D reconvery problem typical in robot application environment. In this study, the imaging system with nonuiformly distributed image sensors simulating human visual system, referred to as Ploar Exponential Grid(PEG), is compared with the existing conventional uniform rectangular grid system in terms of image resolution and computational complexity. By mimicking the geometric structure of the PEG sensor cell, we obtained PEG-like images using computer simulation. With the images obtained from the simulation, image resolution of the two systems are compared and some basic image processing tasks such as image scaling and rotation are implemented based on the PEG sensor system to examine its performance. Furthermore Fourier transform of PEG image is described and implemented in image analysis point of view. Also, the range and heading-angle measurement errors usually encountered in 3-D coordinates recovery with stereo camera system are claculated based on the PEG sensor system and compared with those obtained from the uniform rectangular grid system. In fact, the PEC imaging system not only reduces the computational requirements but also has scale and rotational invariance property in Fourier spectrum. Hence the PEG system has more suitable image coordinate system for image scaling, rotation, and image recognition problem. The range and heading-angle measurement errors with PEG system are less than those of uniform rectangular rectangular grid system in practical measurement range.

• Journal of KIISE
• /
• v.44 no.9
• /
• pp.932-945
• /
• 2017

In this paper, a new hardware and software architecture for a stereo vision processing system including rectification, disparity estimation, and visualization was developed. The developed method, named LArge scale stereo matching method using Support Point Interpolation (LASPI), shows excellence in real-time processing for obtaining dense disparity maps from high quality image regions that contain high density support points. In the real-time processing of high definition (HD) images, LASPI does not degrade the quality level of disparity maps compared to existing stereo-matching methods such as Efficient LArge-scale Stereo matching (ELAS). LASPI has been designed to meet a high frame-rate, accurate distance resolution performance, and a low resource usage even in a limited resource environment. These characteristics enable LASPI to be deployed to safety-critical applications such as an obstacle recognition system and distance detection system for autonomous vehicles. A Field Programmable Gate Array (FPGA) for the LASPI algorithm has been implemented in order to support parallel processing and 4-stage pipelining. From various experiments, it was verified that the developed FPGA system (Xilinx Virtex-7 FPGA, 148.5MHz Clock) is capable of processing 30 HD ($1280{\times}720pixels$) frames per second in real-time while it generates disparity maps that are applicable to real vehicles.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.19 no.1
• /
• pp.66-75
• /
• 2016

Stereoscopic image generated by depth image-based rendering(DIBR) for surveillance robot and camera is appropriate in a low bandwidth network. The image is very important data for the decision-making of a commander and thus its integrity has to be guaranteed. One of the methods used to detect manipulation is to check if the stereoscopic image is taken from the original camera. Sensor pattern noise(SPN) used widely for camera identification cannot be directly applied to a stereoscopic image due to the stereo warping in DIBR. To solve this problem, we find out a shifted object in the stereoscopic image and relocate the object to its orignal location in the center image. Then the similarity between SPNs extracted from the stereoscopic image and the original camera is measured only for the object area. Thus we can determine the source of the camera that was used.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2003.10a
• /
• pp.423-431
• /
• 2003

This study presents on generation coefficients of the RFM using GEO-level stereo images of the IKONOS satellite. 3-D positioning and DEM generation of this model on the test field. In result, the maximum error of image coordinates acquired by the upward transform of the RFM did nat exceed 8 pixels. DEM was generated with kriging interpolation extracted three dimensional ground coordinate to rational quadratic function form, me compared it to reference digital elevation model made from 1:5,000 digital map and 1:1,000 digital map, and so, could generate digital elevation model in the accuracy as average RMSE of elevation was ${\pm}$ 3∼5 m in RFM.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.669-672
• /
• 2003

3D object recognition is difficult but important in computer vision. The important thing is to understand about the relationship between a geometric structure in three dimensions and its image projection. Most 3D recognition systems construct models either manually or by training the pose and orientation of the objects. But both approaches are not satisfactory. In this paper, we focus on a commercial CAD model as a third type of model building for vision. The models are expressed in Initial Graphics Exchanges Specification(IGES) output and reconstructed in a pinhole camera coordinate.

• Proceedings of the IEEK Conference
• /
• 2005.11a
• /
• pp.1261-1264
• /
• 2005

In this paper, a new intelligent moving target tracking and surveillance system basing on the pan/tilt-embedded stereo camera system is suggested and implemented. In the proposed system, once the face area of a target is detected from the input stereo image by using a YCbCr color model and then, using this data as well as the geometric information of the tracking system, the distance and 3D information of the target are effectively extracted in real-time.