• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.11 no.4
• /
• pp.786-792
• /
• 2007

Moving object tracking is very important in video surveillance system. This paper presents a method for tracking moving objects in an outdoor environment. To moving object tracking, first, after extract object that move yielding weight subtraction image and then use close operator to reduce the noise. And we track a object that move detected by matching the extracted multi-feature information. The proposed tracking technique can track moving object by multi-feature matching method so that exactly tracking the objects which are suddenly move or stop. The proposed tracking technique can be efficiently tracking the moving objects, because of combined with spatial position, shape and intensity informations.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.17-20
• /
• 2019

번호판 자동인식 (ALPR: Automatic License Plate Recognition)은 지능형 교통시스템 및 비디오 감시 시스템 등 많은 응용 분야에서 필요한 기술이다. 대부분의 연구는 자동차를 대상으로 번호판 감지 및 인식을 연구하였고, 오토바이를 대상으로 번호판 감지 및 인식은 매우 적은 편이다. 자동차의 경우 번호판이 차량의 전방 또는 후방 중앙에 위치하며 번호판의 뒷배경은 주로 단색으로 덜 복잡한 편이다. 그러나 오토바이의 경우 킥 스탠드를 이용하여 세우기 때문에 주차할 때 오토바이는 다양한 각도로 기울어져 있으므로 번호판의 글자 및 숫자 인식하는 과정이 훨씬 더 복잡하다. 본 논문에서는 다양한 각도로 주차된 오토바이 데이트세트에 대하여 번호판의 문자 인식 정확도를 높이기 위하여 2-스테이지 YOLOv2 알고리즘을 사용하여 오토바이 영역을 선 검출 후 번호판 영역을 검지한다. 인식률을 높이기 위해 앵커박스의 사이즈와 개수를 오토바이 특성에 맞추어 조절하였다. 그 후 기울어진 번호판을 검출한 후 영상 워핑(Image Warping) 알고리즘을 적용하였다. 모의실험 결과, 기존 방식의 인식률이 47,74%에 비해 제안된 방식은 80.23%의 번호판의 인식률을 얻었다. 제안된 방법은 전체적으로 오토바이 번호판 특성에 맞는 앵커박스와 이미지 워핑을 통해서 다양한 기울기의 오토바이 번호판 문자 인식을 높일 수 있었다.

• The Korean Journal of Applied Statistics
• /
• v.35 no.2
• /
• pp.327-333
• /
• 2022

Principal component analysis (PCA) is the most widely used technique in dimension reduction, however, it is very sensitive to outliers. A robust version of PCA, called robust PCA, was suggested by two seminal papers by Candès et al. (2011) and Chandrasekaran et al. (2011). The robust PCA is an essential tool in the artificial intelligence such as background detection, face recognition, ranking, and collaborative filtering. Also, the robust PCA receives a lot of attention in statistics in addition to computer science. In this paper, we introduce recent algorithms for the robust PCA and give some illustrative examples.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.7
• /
• pp.1770-1780
• /
• 1997

Our problem is improvement of image quality because it is inevitable cell loss of image data when traffic congestion occurs. If cells are discarded indiscriminately in transmission of MPEG video data, it occurs severe degradation in quality of service(QOS). In this paper, to solve this problem, we propose two method. The first, we analyze the traffic characteristics of an MPEG encoder and generate high priority and low priority data stream. During network congestion, only the least low priority cells are dropped, and this ensures that the high priority cells are successfully transmitted, which, in turn, guarantees satisfactory QoS. In this case, the prioritization scheme for the encoder assigns components of the data stream to each priority level based on the value of a parameter ${\beta}$. The second, Number of high priority cells are increased when value of ${\beta}$ is large. It occurs the loss of high priority cell in the congestion. To prevent it, this paper is regulated to data stream rate as buffer occupancy with UPC controller. Therefore, encoder's bandwidth can be calculated renegotiation of the encoder and networks. In this paper, the encoder's bandwidth requirements are characterized by a usage parameter control (UPC) set consisting of peak rate, burstness, and sustained rate. An adaptive encoder rate control algorithm at the Networks Interface Card(NIC) computes the necessary UPC parameter to maintain the user specified quality of service. Simulation results are given for a rate-controlled VBR video encoder operating through an ATM network interface which supports dynamic UPC. These results show that dynamic bandwidth renegotiation of prioritized data stream could provided bandwidth saving and significant quality gains which guarantee high priority data stream.

• Journal of Broadcast Engineering
• /
• v.15 no.4
• /
• pp.563-581
• /
• 2010

In this paper, we perform subjective visual quality assessments on UHD video for UHD TV services and analyze the assessment results. Demands for video services have been increased with availabilities of DTV, Internet and personal media equipments. With this trend, the demands for high definition video have also been increasing. Currently, 2K-HD ($1920{\times}1080$) video have been widely consumed over DTV, DVD, digital camcoders, security cameras and other multimedia terminals in various types, and recently digital cinema contents of 4K-UHD($3840{\times}2160$) have been popularly produced and the cameras, beam projects, display panels that support for 4K-UHD video start to come out into multimedia markets. Also it is expected that 4K-UHD service will appear soon in broadcasting and telecommunications environments. Therefore, in this paper, subjective assessments of visual quality on resolutions, color formats, frame rates and compression rates have been carried to provide basis information for standardization of signal specification of UHD video and viewing environments for future UHDTV. As the analysis on the assessments, UHD video exhibits better subjective visual quality than HD by the evaluators. Also, the 4K-UHD test sequences in YUV444 shows better subjective visual quality than the 4K-UHD test sequences in YUV422 and YUV420, but there is little perceptual difference on 4K-UHD test sequences between YUV422 and YUV420 formats. For the comparison between different frame rates, 4K-UHD test sequences of 60fps gives better subjective visual quality than those of 30fps. For bit-depth comparison, HD test sequences in 10-bit depth were little differentiated from those in 8-bit depth in subject visual quality assessment. Lastly, the larger the PSNR values of the reconstructed 4K-UHD test sequences are, the higher the subjective visual quality is. Against the viewing distances, the differences among encoded 4K-UHD test sequences were less distinguished in longer distances from the display.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.47 no.1
• /
• pp.9-16
• /
• 2010

Moving object detection is to detect foreground object different from background scene in a new incoming image frame and is an essential ingredient process in some image processing applications such as intelligent visual surveillance, HCI, object-based video compression and etc. Most of previous object detection algorithms are still computationally heavy so that it is difficult to develop real-time multi-channel moving object detection in a workstation or even one-channel real-time moving object detection in an embedded system using them. Foreground mask correction necessary for a more precise object detection is usually accomplished using morphological operations like opening and closing. Morphological operations are not computationally cheap and moreover, they are difficult to be rendered to run simultaneously with the subsequent connected component labeling routine since they need quite different type of processing from what the connected component labeling does. In this paper, we first devise a fast and precise foreground mask correction algorithm, "Neighbor Foreground Pixel Propagation (NFPP)" which utilizes neighbor pixel checking employed in the connected component labeling. Next, we propose a novel moving object detection method based on the devised foreground mask correction algorithm, NFPP where the connected component labeling routine can be executed simultaneously with the foreground mask correction. Through experiments, it is verified that the proposed moving object detection method shows more precise object detection and more than 4 times faster processing speed for a image frame and videos in the given the experiments than the previous moving object detection method using morphological operations.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.8
• /
• pp.15-22
• /
• 2019

Awireless sensor network utilizes small sensors with a low cost and low power being deployed over a wide area. They monitor the surrounding environment and gather the associated information to transmit it to a base station via multi-hop transmission. Most of the research has mainly focused on static sensors that are located in a fixed position. Unlike a wireless sensor network based on static sensors, we can exploit drone-based technologies for more efficient wireless networks in terms of coverage and energy. In this paper, we introduce a transmission power model and a video encoding power model to design the network environment. We also explain a priority mapping scheme, and deploy drones oriented for network coverage and energy consumption. Through our simulations, this research shows coverage and energy improvements in adrone-based wireless sensor network with fewer sensors, compared to astatic sensor-based wireless sensor network. Concretely, coverage increases by 30% for thedrone-based wireless sensor network with the same number of sensors. Moreover, we save an average of 25% with respect to the total energy consumption of the network while maintaining the coverage required.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.6
• /
• pp.84-90
• /
• 2020

As the uploading and downloading of data through the Internet is becoming more common, data including personal information are easily exposed to unauthorized users. In this study, we detect a target area in images that contain personal information, except for the background, and we protect the detected target area by using a blocking method suitable for the surrounding situation. In this method, only the target area from color image input containing personal information is segmented based on skin color. Subsequently, blurring of the corresponding area is performed in multiple stages based on the surrounding situation to effectively block the detected area, thereby protecting the personal information from being exposed. Experimental results show that the proposed method blocks the object region containing personal information 2.3% more accurately than an existing method. The proposed method is expected to be utilized in fields related to image processing, such as video security, target surveillance, and object covering.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.6
• /
• pp.1209-1214
• /
• 2016

LPR(License plate recognition) is a kind of the most popular surveillance technology based on accompanied by a video and video within the optical character recognition. LPR need a many process. One is a localization of car license plates, license plate of size, space, contrast, normalized to adjust the brightness, another is character division for recognize the character optical character recognition to win the individual characters, character recognition, the other is phrase analysis of the shape, size, position by year, the procedure for the analysis by comparing the database of license plate having a difference by region. In this paper, describing the results of performance of license plate recognition S/W, which was implemented using EmguCV, find the location, using the tesseract OCR, which are well known to an optical character recognition engine of open source, the characters of the license plate image capturing angle of the plate, image size, brightness.

• Journal of Broadcast Engineering
• /
• v.23 no.5
• /
• pp.598-605
• /
• 2018

Behavior awareness is a technology that recognizes human behavior through data and can be used in applications such as risk behavior through video surveillance systems. Conventional behavior recognition algorithms have been performed using the 2D camera image device or multi-mode sensor or multi-view or 3D equipment. When two-dimensional data was used, the recognition rate was low in the behavior recognition of the three-dimensional space, and other methods were difficult due to the complicated equipment configuration and the expensive additional equipment. In this paper, we propose a method of recognizing human behavior using only CCTV images without additional equipment using only RGB and depth information. First, the skeleton extraction algorithm is applied to extract points of joints and body parts. We apply the equations to transform the vector including the displacement vector and the relational vector, and study the continuous vector data through the RNN model. As a result of applying the learned model to various data sets and confirming the accuracy of the behavior recognition, the performance similar to that of the existing algorithm using the 3D information can be verified only by the 2D information.