• The Journal of Korea Robotics Society
• /
• v.14 no.4
• /
• pp.357-362
• /
• 2019

This paper proposes a model and train method that can real-time detect objects and distances estimation based on a monocular camera by applying deep learning. It used YOLOv2 model which is applied to autonomous or robot due to the fast image processing speed. We have changed and learned the loss function so that the YOLOv2 model can detect objects and distances at the same time. The YOLOv2 loss function added a term for learning bounding box values x, y, w, h, and distance values z as 클래스ification losses. In addition, the learning was carried out by multiplying the distance term with parameters for the balance of learning. we trained the model location, recognition by camera and distance data measured by lidar so that we enable the model to estimate distance and objects from a monocular camera, even when the vehicle is going up or down hill. To evaluate the performance of object detection and distance estimation, MAP (Mean Average Precision) and Adjust R square were used and performance was compared with previous research papers. In addition, we compared the original YOLOv2 model FPS (Frame Per Second) for speed measurement with FPS of our model.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.11
• /
• pp.5481-5495
• /
• 2018

Single pixel imaging technology has developed for years, however the video acquisition on the single pixel camera is not a well-studied problem in computer vision. This work proposes a new scheme for single pixel camera to acquire video data and a new regularization for robust signal recovery algorithm. The method establishes a single pixel video compressive sensing scheme to reconstruct the video clips in spatial domain by recovering the difference of the consecutive frames. Different from traditional data acquisition method works in transform domain, the proposed scheme reconstructs the video frames directly in spatial domain. At the same time, a new regularization called spatial cluster is introduced to improve the performance of signal reconstruction. The regularization derives from the observation that the nonzero coefficients often tend to be clustered in the difference of the consecutive video frames. We implement an experiment platform to illustrate the effectiveness of the proposed algorithm. Numerous experiments show the well performance of video acquisition and frame reconstruction on single pixel camera.

• 한국정보통신설비학회:학술대회논문집
• /
• 2008.08a
• /
• pp.372-375
• /
• 2008

IEEE 802.11e standardized the EDCA mechanism to support the priority based QoS. And the virtual collision handler schedules the transmission time of each MAC frame using the internal back-off window according to the access category(AC). This can provides the differentiated QoS to real-time services at the medium traffic load condition. However, the transmission delay of MAC frame for real-time services may be increased as the traffic load of best effort service increases. It becomes more critical when the real-time service uses a compressed mode video codec such as moving picture experts group(MPEG) 4 codec. That is because each frame has the different importance. That is, the I-frame has more information as compared with the P- and the B-frame. In this paper, we proposed a buffer management algorithm based on the frame importance and the delay bound. The proposed algorithm is consisted of the traffic regulator based on the dual token bucket algorithm and the active queue management algorithm. The traffic regulator reduces the transmission rate of lower AC until that the virtual collision handler can transmit an I-frame. And the active queue management discards frame based on the importance of each frame and the delay bound of head of line(HoL) frame when the channel resource is insufficient.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.59 no.8
• /
• pp.1436-1440
• /
• 2010

In this paper, we proposed that stereo camera system using fuzzy control for moving object tracking. We extract some features of the moving object from overall image. This informations send to the PC and the PC calculate the coordinate of the object in the image frame. To make the object set in the middle of the image, the step motor should be controlled accurately and rapidly with the location information received by the PC. Then we design a fuzzy logic system for controlling stereo camera system. To verify the better performance of the proposed algorithm, we exemplified by experimental results.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10b
• /
• pp.412-414
• /
• 2001

In this paper, we describe an real-time facial feature tracker. We only used a general USB PC Camera without a frame grabber. The system has achieved a rate of 8+ frames/second without any low-level library support. It tracks pupils, nostrils and corners of the lip. The signal from USB Camera is YUV 4:2:0 vertical Format. we converted the signal into RGB color model to display the image and We interpolated V channel of the signal to be used for extracting a facial region. and we analysis 2D blob features in the Y channel, the luminance of the image with geometric restriction to locate each facial feature within the detected facial region. Our method is so simple and intuitive that we can make the system work in real-time.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.173.2-173
• /
• 2001

In this paper, an Internet-based monitoring system is studied which analyzes visual image of drifted snow and measures the amount of snowing. The system is composed with reference ruler where LED's are attached, a camera, and PC with a controller. The length of the ruler is 1.5m and the interval of the LED's is 1cm. The controller synchronizes the on-off control of each LED with the image frame of the camera. When the snow is accumulated along the ruler, the corresponding LED is blocked and the lowest position of the lighted LED is estimated with this captured image of the camera. Because of week intensity of LED light and light rays of outdoors, it is hard to discriminate which LED is blocked due to snow. Image processing method is used to obtain proper resolution and good accuracy. This system ...

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1994.04a
• /
• pp.42-48
• /
• 1994

This paper presents the navigation method for a mobile robot which uses a single camera and fluorescent lamp as a guide mark, projected on a convex mirror. The current position and heading direction of the mobile robot are obtained from the image of the guide mark. While the mobile robot travels to a goal position, the current position and heading direction of the mobile robot are updated continuously and the desired path and actual moving path are displayed on the monitor screen in real time. This proposed method eliminates the need to rotate the camera to track the guide mark, since a panoramic view of the surrounding area is available from the convex mirror, and natural guide marks such as usual florescent lamp on the ceiling or door frame can be used for navigation.

• Fashion & Textile Research Journal
• /
• v.4 no.3
• /
• pp.284-288
• /
• 2002

In this research, a new quantitative fabric drape evaluation system has been developed using image processing technology. The purpose of this research is to get the more detailed information of fabric drapability quantitatively from digital images captured with a digital camera generally commercialized. The shape parameters of a 3-dimensional geometric drape model were defined as the number of nodes, frequency and amplitude. Also, various statistical information of drape shapes can be obtained using image processing technology and frequency analysis as well as traditional drape coefficients. Hardware system to capture drape images is simply composed of three parts including a digital USB (Universal Serial Bus) camera, a frame cover and a stand for camera to attach to traditional drape tester. The evaluation software coded with the MS Visual C++ is operated under the MS windows 9x above.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2009.06a
• /
• pp.703-704
• /
• 2009

• Architectural research
• /
• v.10 no.1
• /
• pp.33-39
• /
• 2008

Condensation has mainly occurred in corridor type apartment door which is exposed to the outside air and is made of steel, which has high thermal conductivity. As a result, the total costs of repair have increased with the number of disputes with residents. In this study, therefore, we investigate materials and construction methods used in apartment door, perform a computer simulation to find out possible improvements, and then suggest the dew point to prevent the occurrence of condensation throughout simulation. The results indicate that the temperature that condensation does not occur is $15.4^{\circ}C$, and the optimum method of achieving this dew point is shown to be a door frame system including a large vertical slot to decrease the area of thermal conduction between the outer and inner portions of the door frame. Mock-up tests show that the surface temperature of the door frame was higher than the dew point, and the system can withstand severe cold conditions of $-20^{\circ}C$. In application test, the surface temperature of door frame with vertical slots is $5.9^{\circ}C$in average, which is higher than the existing door frame. Furthermore, in the temperature distribution of the surrounding door measured with infrared ray camera, the existing door shows the high temperature distribution indicating lack of insulation, but the improved door shows the low temperature distribution indicating higher insulation.