• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.6
• /
• pp.1263-1268
• /
• 2018

In this paper, a video analysis is proposed to implement video surveillance system with deep learning object detection and probabilistic data association filter for tracking multiple objects, and suggests its implementation using GPU. The proposed video analysis technique involves object detection and object tracking sequentially. The deep learning network architecture uses ResNet for object detection and applies probabilistic data association filter for multiple objects tracking. The proposed video analysis technique can be used to detect intruders illegally trespassing any restricted area or to count the number of people entering a specified area. As a results of simulations and experiments, 48 channels of videos can be analyzed at a speed of about 27 fps and real-time video analysis is possible through RTSP protocol.

• Journal of information and communication convergence engineering
• /
• v.12 no.4
• /
• pp.263-270
• /
• 2014

Corner detection and feature extraction are essential aspects of computer vision problems such as object recognition and tracking. Feature detectors such as Scale Invariant Feature Transform (SIFT) yields high quality features but computationally intensive for use in real-time applications. The Features from Accelerated Segment Test (FAST) detector provides faster feature computation by extracting only corner information in recognising an object. In this paper we have analyzed the efficient object detection algorithms with respect to efficiency, quality and robustness by comparing characteristics of image detectors for corner detector and feature extractors. The simulated result shows that compared to conventional SIFT algorithm, the object recognition system based on the FAST corner detector yields increased speed and low performance degradation. The average time to find keypoints in SIFT method is about 0.116 seconds for extracting 2169 keypoints. Similarly the average time to find corner points was 0.651 seconds for detecting 1714 keypoints in FAST methods at threshold 30. Thus the FAST method detects corner points faster with better quality images for object recognition.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.05a
• /
• pp.671-674
• /
• 2018

In this study, we implement a YOLO driving assistance system using a model car. The YOLO is an object detection and recognition algorithm using deep running which is becoming an issue recently. The system alerts the lane departure by applying the image processing technology to the image acquired through the camera, recognizes the objects using the YOLO, and performs various functions according to the type of the object and the distance between the vehicle. the YOLO, which is superior to the existing object detection and recognition algorithm, improves the performance of the driving assist system without additional equipment. The driving assist system using the YOLO will ensure the safety of the driver with low cost.

• Journal of the Optical Society of Korea
• /
• v.8 no.4
• /
• pp.174-181
• /
• 2004

Various kinds of systems, that can do target recognition and position detection simultaneously by using infrared sensing detectors, have been developed. In this paper, the detection system TRSS (Thermal target Recognition by Spiral Scanning) adopts linear array shaped uncooled IR detector and uses spiral type fast scanning method for relative position detection of target objects, which radiate an IR region wavelength spectrum. It can detect thermal energy radiating from a 9 m-size target object as far as 200 m distance. And the maximum field of a detector is fully filled with the same size of target object at the minimum approaching distance 50 m. We investigate two types of lens systems. One is a singlet lens and the other is a doublet lens system. Every system includes one aspheric surface and free positioned aperture stop. Many designs of F/1.5 system with ${\pm}5.2^{\circ}$ field at the Efl=20, 30 mm conditions for single element and double elements lens system respectively are compared in their resolution performance [MTF] according to the aspheric surface and stop position changing on their optimization process. Optimum design is established including mechanical boundary conditions and manufacturing considerations.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.4
• /
• pp.93-99
• /
• 2020

It seems to be increased rapidly that practical uses of intelligent Dron for public mission performance such as surveillance, prevention of disaster accident, relief etc with Dron technology development. Dron is needed for major technology realization of detection and trace technology of target, flight control and obstacle avoidance during flighting, detection and control of landing point functions to use smart safety-city platform construction. This dron system cause a great ripple effect technically and promote industrialization in the field of new technology. In this paper, an effective analysis and design method of dron system software will be presented by showing user requirement analysis using object-oriented method, flowchart and screen design.

• Journal of Software Assessment and Valuation
• /
• v.16 no.2
• /
• pp.99-107
• /
• 2020

An underground integrated map is being built for underground safety management and is being updated periodically. The map update proceeds with the procedure of deleting all previously stored objects and saving newly entered objects. However, even unchanged objects are repeatedly stored, deleted, and stored. That causes the delay of the update time. In this study, in order to shorten the update time of the integrated map, an updated object and an unupdated object are separated, and only updated objects are reflected in the underground integrated map, and a system implementing this technology is described. For the updated object, an object comparison method using the center point of the object is used, and a quad tree is used to improve the search speed. The types of updated objects are classified into addition and deletion using the shape of the object, and change using its attributes. The proposed system consists of update object detection, extraction, conversion, storage, and history management modules. This system has the advantage of being able to update the integrated map about four times faster than the existing method based on the data used in the experiment, and has the advantage that it can be applied to both ground and underground facilities.

• Journal of the Semiconductor & Display Technology
• /
• v.23 no.1
• /
• pp.19-24
• /
• 2024

It is not an exaggeration to say that a single fire causes a lot of damage, so fires are one of the disaster situations that must be alerted as soon as possible. Various technologies have been utilized so far because preventing and detecting fires can never be completely accomplished with individual human efforts. Recently, deep learning technology has been developed, and fire detection systems using object detection neural networks are being actively studied. In this paper, we propose a new fire detection system that improves the previously studied fire detection system. We train the YOLOv8 model using refined datasets through improved labeling methods, derive results, and demonstrate the superiority of the proposed system by comparing it with the results of previous studies.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.1141-1142
• /
• 2008

For intelligent vehicles, recognizing the sounding is an important task. In this paper we propose an road area detection system. This system uses u-disparity and v-disparity map. v-disparity map is used to find the road area. u-disparity is used to cluster the area that is an object. The test results and overall system is discribed in this paper.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2018.01a
• /
• pp.135-136
• /
• 2018

본 논문에서는 산업현장에서 특정한 물건을 인식하고 판단하여 로봇팔로 운반할 수 있는 딥러닝을 적용한 객체 인식 기반의 로봇 팔 제어 시스템을 제안하였다. 제안한 시스템은 깊이 인식 카메라를 이용하여 3D 이미지를 촬영 하고 딥러닝으로 검출된 객체를 판별 및 분류 후 인식된 객체를 로봇 팔로 피킹 하도록 구현하였다. 이를 통해, 딥러닝과 깊이인식 카메라로 다양한 환경에서 객체를 정확히 분류 및 추적할 수 있도록 해서 스마트팩토리등 다양한 분야에 활용할 수 있는 시스템을 제안하였다.

• Journal of the Korea Society of Computer and Information
• /
• v.22 no.12
• /
• pp.79-85
• /
• 2017

In this paper, we propose a CNN-based swimmer detection algorithm. Every year, water safety accidents have been occurred frequently, and accordingly, intelligent video surveillance systems are being developed to prevent accidents. Intelligent video surveillance system is a real-time system that detects objects which users want to do. It classifies or detects objects in real-time using algorithms such as GMM (Gaussian Mixture Model), HOG (Histogram of Oriented Gradients), and SVM (Support Vector Machine). However, HOG has a problem that it cannot accurately detect the swimmer in a complex and dynamic environment such as a beach. In other words, there are many false positives that detect swimmers as waves and false negatives that detect waves as swimmers. To solve this problem, in this paper, we propose a swimmer detection algorithm using CNN (Convolutional Neural Network), specialized for small object sizes, in order to detect dynamic objects and swimmers more accurately and efficiently in complex environment. The proposed CNN sets the size of the input image and the size of the filter used in the convolution operation according to the size of objects. In addition, the aspect ratio of the input is adjusted according to the ratio of detected objects. As a result, experimental results show that the proposed CNN-based swimmer detection method performs better than conventional techniques.