• International Journal of Aeronautical and Space Sciences
• /
• 제12권3호
• /
• pp.252-259
• /
• 2011

This paper is focused on dynamic modeling and control system design as well as vision based collision avoidance for multi-rotor unmanned aerial vehicles (UAVs). Multi-rotor UAVs are defined as rotary-winged UAVs with multiple rotors. These multi-rotor UAVs can be utilized in various military situations such as surveillance and reconnaissance. They can also be used for obtaining visual information from steep terrains or disaster sites. In this paper, a quad-rotor model is introduced as well as its control system, which is designed based on a proportional-integral-derivative controller and vision-based collision avoidance control system. Additionally, in order for a UAV to navigate safely in areas such as buildings and offices with a number of obstacles, there must be a collision avoidance algorithm installed in the UAV's hardware, which should include the detection of obstacles, avoidance maneuvering, etc. In this paper, the optical flow method, one of the vision-based collision avoidance techniques, is introduced, and multi-rotor UAV's collision avoidance simulations are described in various virtual environments in order to demonstrate its avoidance performance.

• 전기학회논문지
• /
• 제67권9호
• /
• pp.1224-1231
• /
• 2018

Automatic age estimation has been used in many social network applications, practical commercial applications, and human-computer interaction visual-surveillance biometrics. However, it has rarely been explored. In this paper, we propose an automatic age estimation system, which includes face detection and convolutional deep learning based on an inception module. The latter is a 22-layer-deep network that serves as the particular category of the inception design. To evaluate the proposed approach, we use 4,000 images of eight different age groups from the Adience age dataset. k-fold cross-validation (k = 5) is applied. A comparison of the performance of the proposed work and recent related methods is presented. The results show that the proposed method significantly outperforms existing methods in terms of the exact accuracy and off-by-one accuracy. The off-by-one accuracy is when the result is off by one adjacent age label to the above or below. For the exact accuracy, the age label of "60+" is classified with the highest accuracy of 76%.

• Journal of Information Processing Systems
• /
• 제13권5호
• /
• pp.1299-1319
• /
• 2017

This paper presents a scalable multiple camera collaboration strategy for active tracking applications in large areas. The proposed approach is based on distributed mechanism but emulates the master-slave mechanism. The master and slave cameras are not designated but adaptively determined depending on the object dynamic and density distribution. Moreover, the number of cameras emulating the master is not fixed. The collaboration among the cameras utilizes global and local sectors in which the visual correspondences among different cameras are determined. The proposed method combines the local information to construct the global information for emulating the master-slave operations. Based on the global information, the load balancing of active tracking operations is performed to maximize active tracking coverage of the highly dynamic objects. The dynamics of all objects visible in the local camera views are estimated for effective coverage scheduling of the cameras. The active tracking synchronization timing information is chosen to maximize the overall monitoring time for general surveillance operations while minimizing the active tracking miss. The real-time simulation result demonstrates the effectiveness of the proposed method.

• 스마트미디어저널
• /
• 제11권8호
• /
• pp.84-92
• /
• 2022

Precision livestock monitoring promises greater management efficiency for farmers and higher welfare standards for animals. Recent studies on video-based animal activity recognition and tracking have shown promising solutions for understanding animal behavior. To achieve that, surveillance cameras are installed diagonally above the barn in a typical cattle farm setup to monitor animals constantly. Under these circumstances, tracking individuals requires addressing challenges such as occlusion and visual appearance, which are the main reasons for track breakage and increased misidentification of animals. This paper presents a framework for multi-cattle tracking in closed barns with appearance and motion models. To overcome the above challenges, we modify the DeepSORT algorithm to achieve higher tracking accuracy by three contributions. First, we reduce the weight of appearance information. Second, we use an Ensemble Kalman Filter to predict the random motion information of cattle. Third, we propose a supplementary matching algorithm that compares the absolute cattle position in the barn to reassign lost tracks. The main idea of the matching algorithm assumes that the number of cattle is fixed in the barn, so the edge of the barn is where new trajectories are most likely to emerge. Experimental results are performed on our dataset collected on two cattle farms. Our algorithm achieves 70.37%, 77.39%, and 81.74% performance on HOTA, AssA, and IDF1, representing an improvement of 1.53%, 4.17%, and 0.96%, respectively, compared to the original method.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권22호
• /
• pp.9731-9737
• /
• 2014

Large scale secular registry or surveillance systems have been accumulating vast data that allow mathematical modeling of cancer incidence and mortality rates. Most contemporary models in this regard use time series and APC (age-period-cohort) methods and focus primarily on predicting or analyzing cancer epidemiology with little attention being paid to implications for designing cancer registry, surveillance or evaluation initiatives. This research models age-specific cancer incidence rates using logistic growth equations and explores their performance under different scenarios of data completeness in the hope of deriving clues for reshaping relevant data collection. The study used China Cancer Registry Report 2012 as the data source. It employed 3-parameter logistic growth equations and modeled the age-specific incidence rates of all and the top 10 cancers presented in the registry report. The study performed 3 types of modeling, namely full age-span by fitting, multiple 5-year-segment fitting and single-segment fitting. Measurement of model performance adopted adjusted goodness of fit that combines sum of squred residuals and relative errors. Both model simulation and performance evalation utilized self-developed algorithms programed using C# languade and MS Visual Studio 2008. For models built upon full age-span data, predicted age-specific cancer incidence rates fitted very well with observed values for most (except cervical and breast) cancers with estimated goodness of fit (Rs) being over 0.96. When a given cancer is concerned, the R valuae of the logistic growth model derived using observed data from urban residents was greater than or at least equal to that of the same model built on data from rural people. For models based on multiple-5-year-segment data, the Rs remained fairly high (over 0.89) until 3-fourths of the data segments were excluded. For models using a fixed length single-segment of observed data, the older the age covered by the corresponding data segment, the higher the resulting Rs. Logistic growth models describe age-specific incidence rates perfectly for most cancers and may be used to inform data collection for purposes of monitoring and analyzing cancer epidemic. Helped by appropriate logistic growth equations, the work vomume of contemporary data collection, e.g., cancer registry and surveilance systems, may be reduced substantially.

• 한국멀티미디어학회논문지
• /
• 제11권3호
• /
• pp.301-314
• /
• 2008

기존의 이벤트 인식은 한정된 규칙 기반으로 이루어졌고, 시나리오 해석은 확률 자료의 크기로 많은 학습 시간이 필요했다. 본 논문에서는 객체로부터 특징 벡터를 추출하고 각 객체의 행동 양식을 분석하여 현재 객체의 이벤트를 인식하는 방법과 확률 모델을 기반으로 본 논문에서 정의한 긴급 상황 시나리오를 해석할 수 있는 방법을 제안한다. 독립 객체의 이벤트 규칙은 주-이벤트, 움직임-이벤트, 상호-이벤트, 그리고 'FALL DOWN' 이벤트로 구성되며, 객체의 특징 벡터와 베이지안 네트워크에 의해 학습된 분할 움직임 방향 벡터(SMOV)를 통해 정의된다. 긴급 상황 시나리오는 현재 이벤트의 상태와 사후 확률에 의해 분석된다. 본 논문에서는 기존 방법에 비해 다양한 이벤트를 정의하였고 이벤트 간의 독립성을 높여 확장성이 용이하도록 하였다. 그리고 객체 추적만을 통해 얻을 수 없는 의미론적 정보를 규칙과 확률을 기반으로 획득할 수 있었다.

• 한국산학기술학회논문지
• /
• 제16권3호
• /
• pp.2117-2122
• /
• 2015

한 장의 영상에 보다 많은 영상정보를 담기 위해서는 넓은 시야각이 필요하다. 넓은 시야각을 갖는 영상은 보안, 감시, 원격화상회의, 이동로봇 등의 산업분야에서 사용된다. 본 논문에서는 광각의 파노라마 영상을 획득하기 위해 쌍곡면 실린더형 반사체를 이용한 영상획득 방법을 제안한다. 일반적인 응용 예에서 수직화각 보다 수평화각이 중요하므로 수직방향으로는 평면거울과 같고, 수평방향으로 쌍곡선 형태를 갖는 실린더형 반사체를 설계하였다. 광학적 성능 분석을 위해 광선추적법을 통해 본 쌍곡면 실린더형 반사체 영상계의 영상획득 모델을 구하였으며, 쌍곡면 실린더형 반사체를 실제 제작하였고, 영상 실험을 통해 광각 영상획득 성능을 검증하였다. 제안하는 영상 시스템은 기존 방법에 비해 경제적이며, 별도의 영상처리 없이 수평화각 210도에 이르는 광각의 실시간 파노라마 영상을 획득할 수 있었다.

• 한국동물위생학회지
• /
• 제40권1호
• /
• pp.15-20
• /
• 2017

A two-tube reverse transcription loop-mediated isothermal amplification (RT-LAMP) assay was designed for the rapid visual detection of the M gene of all subtypes of avian influenza virus (AIV) and the H5 gene of the H5 subtype of highly pathogenic AIV (HPAIV). The reaction carried out in two tubes in a single step at $58^{\circ}C$ for 40 min, and the assay results could be visually detected by using hydroxynaphthol blue dye. Using M or H5 gene-specific primers, the assay successfully detected all subtypes or H5 subtypes of AIVs, including the Korean representative H5N1 and H5N8 HPAIVs. The detection limit of the assay was approximately $10^{2.0}$ $EID_{50}/reaction$ for the M and H5 genes of H5N1 HPAIV, respectively, and was more sensitive than that of previously reported RT-LAMP and comparable to that of real-time RT-PCR. These results suggest that the present RT-LAMP assay, with its high specificity, sensitivity, and simplicity, will be a useful diagnostic tool for surveillance of currently circulating H5 HPAIVs and other subtypes of AIV in bird population, even in under-equipped laboratories.

• 제어로봇시스템학회논문지
• /
• 제14권8호
• /
• pp.848-855
• /
• 2008

In this paper, we describe correspondence among multiple images taken by multiple cameras. The correspondence among multiple views is an interesting problem which often appears in the application like visual surveillance or gesture recognition system. We use the principal axis and the ground plane homography to estimate foot of human. The principal axis belongs to the subtracted silhouette-based region of human using subtraction of the predetermined multiple background models with current image which includes moving person. For the calculation of the ground plane homography, we use landmarks on the ground plane in 3D space. Thus the ground plane homography means the relation of two common points in different views. In the normal human being, the foot of human has an exactly same position in the 3D space and we represent it to the intersection in this paper. The intersection occurs when the principal axis in an image crosses to the transformed ground plane from other image. However the positions of the intersection are different depend on camera views. Therefore we construct the correspondence that means the relationship between the intersection in current image and the transformed intersection from other image by homography. Those correspondences should confirm within a short distance measuring in the top viewed plane. Thus, we track a person by these corresponding points on the ground plane. Experimental result shows the accuracy of the proposed algorithm has almost 90% of detecting person for tracking based on correspondence of intersections.

• 전자공학회논문지SC
• /
• 제48권5호
• /
• pp.45-51
• /
• 2011

본 논문에서는 비교사학습법을 통해 영상의 방대한 정보를 효율적으로 모델링 하는 방법을 제안하고자 한다. 여기서 이동궤적들은 자연어 처리에 사용되는 알고리즘인 잠재 디리클레 할당 모형(Latent Dirichlet Allocation)에 의해 직진, 좌회전, 우회전등 각 상황 별로 주제에 따라 그 영역을 효과적으로 분류할 수 있다. LDA를 이용해 주제별로 의미 있는 영역을 분류한 후, 각 주제별로 분류된 궤적을 관측열로 보고 은닉 마르코프 모델(Hidden Markov Model)의 바움-웰치 알고리즘을 사용하여 학습한다. 전향 알고리즘을 사용하여 입력된 행동과 학습된 행동을 비교함으로써 영상내의 행동이 정상인지 비정상인지를 효과적으로 판단할 수 있다. 실험결과 다양한 영상에 대해 의미있는 주제별로 영역이 잘 분류되며 추적에러로 인한 궤적의 노이즈에도 강인하게 물체의 무단횡단, 신호위반과 같은 상황을 효과적으로 탐지하는 것을 확인할 수 있다.