• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.14-25
• /
• 2021

Change detection, which finds changes in remote sensing images of the same location captured at different times, is very important because it is used in various applications. However, registration errors, building displacement errors, and shadow errors cause false positives. To solve these problems, we propose a novle deep convolutional network called CADNet (Change Attention Dense Siamese Network). CADNet uses FPN (Feature Pyramid Network) to detect multi-scale changes, applies a Change Attention Module that attends to the changes, and uses DenseNet as a feature extractor to use feature maps that contain both low-level and high-level features for change detection. CADNet performance measured from the Precision, Recall, F1 side is 98.44%, 98.47%, 98.46% for WHU datasets and 90.72%, 91.89%, 91.30% for LEVIR-CD datasets. The results of this experiment show that CADNet can offer better performance than any other traditional change detection method.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.14 no.3
• /
• pp.51-56
• /
• 2014

TETRA(TErrestrial Trunked RAdio) provide specialized disaster radio communication services as a standard European for digital TRS(Trunked Radio System). Especially, DMO(Direct Mode Operation) feature is used effectively in the radio shadow areas which base station does not propagate radio signal because it can communicate directly with terminal to terminal without base station's relay function. However, to effectively used DMO feature, radio coverage prediction information should be provided to users. Therefore in this paper, we were calculated link budget of TETRA DMO terminals which were distributing and operating in the country and then predicted reaching distance about radio propagation to be applied with path loss model.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.06a
• /
• pp.194-197
• /
• 2016

자동차 산업이 발전하면서 안정적인 주행과 운전자의 편의성을 위한 지능형운전자보조시스템인 ADAS (Advanced Driver Assistance System)가 이슈가 되고 있다. 차선인식의 결과에 따라 차선이탈 경고시스템의 성능이 달라지기 때문에 차선인식은 ADAS에서 매우 중요한 핵심적인 기술이라 할 수 있다. 이에 본 논문에서는 그림자 영역과 같이 밝기의 분포가 균일하지 않는 환경에서 강인하게 동작하는 차선인식 알고리즘을 제안하였다, 지역적인 밝기 특징을 고려하여 차선에 해당하는 특징점을 추출하며, 추출된 특징점 가운데 이상치(outlier)를 제거하기 위해 RANSAC (RANdom SAmple Consensus) 알고리즘을 이용하여 차선을 검출한다. 또한 RANSAC 알고리즘에서 신뢰도가 높은 차선이 검출되면 그 주위에 특징점을 추출하기 위한 관심영역을 설정함으로써 안정적인 차선 검출이 가능하도록 하였다.

• Journal of ILASS-Korea
• /
• v.11 no.2
• /
• pp.105-112
• /
• 2006

A unique feature of fluidic atomizers is that the nozzle geometry produces a thin capillary Jet which is forced to oscillate on a 2-dimensional plane through the use of a passive feedback mechanism. The objective of the current work is to characterize the influence of the stagnation pressure at the nozzle exit, jet oscillation and stretching on the breakup properties of the capillary ligament. To achieve this, shadow graph technique is used to measure size, shape, velocity and the number density of the droplets as a function of the position within the spray fan. The breakup length, defined as the radial distance from the breakup point, is analyzed as a function of the non-dimensional parameters. Finally, a kinematic model is developed to simulate the breakup of the oscillating jets at low stagnation pressures. Using the existing jet breakup theories, the model is used to predict the size and diameter distribution of the droplets after primary atomization.

• Journal of Korea Multimedia Society
• /
• v.20 no.2
• /
• pp.144-152
• /
• 2017

Among the technologies for implementing autonomous vehicles, advanced driver assistance system is a key technology to support driver's safe driving. In the technology using the vision sensor having a high utility, various preprocessing methods are used prior to feature extraction for lane detection. However, in the existing methods, the unnecessary lane candidates such as cars, lawns, and road separator in the road area are false positive. In addition, there are cases where the lane candidate itself can not be extracted in the area under the overpass, the lane within the dark shadow, the center lane of yellow, and weak lane. In this paper, we propose an efficient preprocessing method using k-means clustering for image division and the HSV color model. When the proposed preprocessing method is applied, the true positive region is maximally maintained during the lane detection and many false positive regions are removed.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.11a
• /
• pp.347-350
• /
• 2011

This paper propose a statistical color model of background extraction base on Hue-Saturation-Value(HSV) color space, instead of the traditional RGB space, and shows that it provides a better use of the color information. HSV color space corresponds closely to the human perception of color and it has revealed more accuracy to distinguish shadows [3] [4]. The key feature of this segmentation method is based on processing hue component of color in HSV color space on image area. The HSV color model is used, its color components are efficiently analyzed and treated separately so that the proposed algorithm can adapt to different environmental illumination condition and shadows. Polar and linear statistical operations are used to calculate the background from the video frames. The experimental results show that the proposed background subtraction method can automatically segment video objects robustly and accurately in various illuminating and shadow environments.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.44 no.6
• /
• pp.84-92
• /
• 2007

The gait recognition is presented for human identification from a sequence of noisy silhouettes segmented from video by capturing at a distance. The proposed gait recognition algorithm gives better performance than the baseline algorithm because of segmentation of the object by using multiple modules; i) motion detection, ii) object region detection, iii) head detection, and iv) active shape models, which solve the baseline algorithm#s problems to make background, to remove shadow, and to be better recognition rates. For the experiment, we used the HumanID Gait Challenge data set, which is the largest gait benchmarking data set with 122 objects, For realistic simulation we use various values for the following parameters; i) viewpoint, ii) shoe, iii) surface, iv) carrying condition, and v) time.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2007.05a
• /
• pp.1053-1056
• /
• 2007

The acoustical performance of noise reducing devices installed on the top of a noise barrier were tested by small-scale outdoor tests. Noise measurements before and after installation of the devices were carried out using sound intensity methods. It is well known that the sound intensity method can specify the strength and directivity of sound, and it is convenient to consider the feature of sound around a noise barrier. The noise reduction effect of each edge device was evaluated using the difference between the input and output sound power levels calculated from sound intensities. It was investigated that each device had different efficiency in the shadow zone, while there was no significant difference between edge devices in the illuminated zone.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.9
• /
• pp.95-103
• /
• 1998

This paper proposes an image segmentation algorithm based on fusion of range and intensity images. Based on Bayesian theory, a priori knowledge is encoded by the Markov random field (MRF). A maximum a posteriori (MAP) estimator is constructed using the features extracted from range and intensity images. Objects are approximated by local planar surfaces in range images, and the parametric space is constructed with the surface parameters estimated pixelwise. In intensity images the ${\alpha}$-trimmed variance constructs the intensity feature. An image is segmented by optimizing the MAP estimator that is constructed using a likelihood function based on edge information. Computer simulation results shw that the proposed fusion algorithm effectively segments the images independentl of shadow, noise, and light-blurring.

• Publications of The Korean Astronomical Society
• /
• v.26 no.3
• /
• pp.129-140
• /
• 2011

In this paper, we have studied Sogyupyo (小圭表, small noon gnomon) of the Joseon dynasty. According to the Veritable Records of King Sejong (世宗, 1418 - 1450), Daegyupyo (大圭表, large noon gnomon) with a height of 40-feet [尺] was constructed by Jeong, Cho (鄭招) and his colleagues in 1435, and installed around Ganuidae (簡儀臺, platform of Ganui). On the contrary, the details of Sogyupyo are unknown although the shadow length measurements by Daegyupyo and Sogyupyo are found on the Veritable Records of King Myeongjong (明宗, 1545 - 1567). By analysing historical documents and performing experiments, we have investigated the construction details of Sogyupyo including its development year, manufacturer, and installation spot. We have found that Sogyupyo would be manufactured by King Sejong in 1440 and placed around Ganuidae. And Sogyupyo would be five times smaller than Daegyupyo, i.e., 8-feet. On the basis of experiments, we suggest that although it is smaller, Sogyupyo was equipped with a bar [橫梁] and a pin-hole projector [影符] like Daegyupyo in order to produce the observation precision presented in the Veritable Record of King Myeongjong.