• The Journal of the Korea institute of electronic communication sciences
• /
• v.9 no.2
• /
• pp.155-160
• /
• 2014

In this paper, an algorithm using RGB-depth camera is proposed to detect smoke in interrior. RGB-depth camera, the Kinect provides RGB color image and depth information. The Kinect sensor consists of an infra-red laser emitter, infra-red camera and an RGB camera. A specific pattern of speckles radiated from the laser source is projected onto the scene. This pattern is captured by the infra-red camera and is analyzed to get depth information. The distance of each speckle of the specific pattern is measured and the depth of object is estimated. As the depth of object is highly changed, the depth of object plain can not be determined by the Kinect. The depth of smoke can not be determined too because the density of smoke is changed with constant frequency and intensity of infra-red image is varied between each pixels. In this paper, a smoke detection algorithm using characteristics of the Kinect is proposed. The region that the depth information is not determined sets the candidate region of smoke. If the intensity of the candidate region of color image is larger than a threshold, the region is confirmed as smoke region. As results of simulations, it is shown that the proposed method is effective to detect smoke in interior.

• Journal of KIISE
• /
• v.41 no.11
• /
• pp.927-934
• /
• 2014

In this paper, we propose an effective method for recognizing daily human activities from a stream of three dimensional body poses, which can be obtained by using Kinect-like RGB-D sensors. The body pose data provided by Kinect SDK or OpenNI may suffer from both the view variance problem and the scale variance problem, since they are represented in the 3D Cartesian coordinate system, the origin of which is located on the center of Kinect. In order to resolve the problem and get the view-invariant and scale-invariant features, we transform the pose data into the spherical coordinate system of which the origin is placed on the center of the subject's hip, and then perform on them the scale normalization using the length of the subject's arm. In order to represent effectively complex internal structures of high-level daily activities, we utilize Hidden state Conditional Random Field (HCRF), which is one of probabilistic graphical models. Through various experiments using two different datasets, KAD-70 and CAD-60, we showed the high performance of our method and the implementation system.

• Smart Media Journal
• /
• v.1 no.3
• /
• pp.29-35
• /
• 2012

A depth image-based rendering (DIBR) technique is one of the rendering processes of virtual views with a color image and the corresponding depth map. The most important issue of DIBR is that the virtual view has no information at newly exposed areas, so called dis-occlusion. In this paper, we propose an intermediate view generation algorithm using the Kinect depth camera that utilizes the infrared structured light. After we capture a color image and its corresponding depth map, we pre-process the depth map. The pre-processed depth map is warped to the virtual viewpoint and filtered by median filtering to reduce the truncation error. Then, the color image is back-projected to the virtual viewpoint using the warped depth map. In order to fill out the remaining holes caused by dis-occlusion, we perform a background-based image in-painting operation. Finally, we obtain the synthesized image without any dis-occlusion. From experimental results, we have shown that the proposed algorithm generated very natural images in real-time.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.2
• /
• pp.439-446
• /
• 2017

Portable low-cost Kinect sensor was used to analyze standing balance ability of the elderly. Eighty subjects who can walk alone and have a normal cognitive level participated in this experiment. Based on Berg Balance scale (BBS) test with 52 points, subjects were divided into Healthy older (HO: 46 persons, BBS: $53.80{\pm}1.19$) and Impaired older (IO: 34 persons, BBS: $49.06{\pm}2.03$) group. Each subject performed 30 seconds four different standing balance tests (EO: Eyes Open, EC: Eyes Close, EOf: Eyes Open on foam, ECf: Eyes Close on foam). Five variables (Mean distance, Range of distance, Root mean square, Mean velocity, 95% ellipse area) were calculated from the hip joint center movement of Kinect sensor. Results showed that there were significant differences between groups for four different standing tests. Calculated variables from kinect sensor showed significant correlation with BBS score. Especially, mediolateral mean distance, mediolateral root mean square, mediolateral range of distance and 95% ellipse area showed discriminative ability for all tests. Mean values of variables of IO were higher than those of HO, which means the decreased balance ability in IO compared with HO. Therefore, it was possible to estimate simple balance assessment of the elderly using portable low-cost Kinect sensor.

• Journal of information and communication convergence engineering
• /
• v.17 no.1
• /
• pp.74-83
• /
• 2019

This study designs a squat posture recognition system that can provide correct squat posture guidelines. This system comprises two modules: a Kinect camera for monitoring users' body movements and a Wii Balance Board(WBB) for measuring balanced postures with legs. Squat posture recognition involves two states: "Stand" and "Squat." Further, each state is divided into two postures: correct and incorrect. The incorrect postures of the Stand and Squat states were classified into three and two different types of postures, respectively. The factors that determine whether a posture is incorrect or correct include the difference between shoulder width and ankle width, knee angle, and coordinate of center of pressure(CoP). An expert and 10 participants participated in experiments, and the three factors used to determine the posture were measured using both Kinect and WBB. The acquired data from each device show that the expert's posture is more stable than that of the subjects. This data was classified using a support vector machine (SVM) and $na{\ddot{i}}ve$ Bayes classifier. The classification results showed that the accuracy achieved using the SVM and $na{\ddot{i}}ve$ Bayes classifier was 95.61% and 81.82%, respectively. Therefore, the developed system that used Kinect and WBB could classify correct and incorrect postures with high accuracy. Unlike in other studies, we obtained the spatial coordinates using Kinect and measured the length of the body. The balance of the body was measured using CoP coordinates obtained from the WBB, and meaningful results were obtained from the measured values. Finally, the developed system can help people analyze the squat posture easily and conveniently anywhere and can help present correct squat posture guidelines. By using this system, users can easily analyze the squat posture in daily life and suggest safe and accurate postures.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.11
• /
• pp.2095-2102
• /
• 2017

In this paper, we implemented the elderly home training contents provide individual exercise prescription according to the user's athletic ability and provide personalized program to the elderly individual. Health promotion is essential for overcoming the low health longevity of senior citizens preparing for aging population. Therefore, the lower body strengthening exercise to prevent falls is crucial to prevent a fall in the number of deaths of senior citizens. In this game model, the elderly are aiming at home training contents that can be found to feel that the elderly are going out of walk and exercising in the natural environment. To achieve this, Kinect extracts a specific bone model provide by the Kinect Sensor to generate the feature vectors and recognizes the movements and motion of the user. The recognition test using the Kinect sensor showed a recognition rate of about 80 to 97%.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2012.05a
• /
• pp.182-184
• /
• 2012

마이크로소프트에서 개발한 키넥트(kinect)는 엑스박스(XBox) 게임 컨트롤러로 사용하는 장비이며 이 센서를 이용하여 사용자의 인체 행동을 인식하여 게임을 진행할 수 있는 센서 시스템이다. 또한 윈도우 환경에서 키넥트를 활용하여 다양한 응용 프로그램 개발을 할 수 있도록 SDK를 제공하고 있다. 현대사회에서 각종 범죄가 늘어남에 따라서 CCTV의 운용이 늘어나고 있으며 지정된 구역을 감시하는데 다양한 영상 장비들과 프로그램이 운용하고 있다. 시장에 판매되고 있는 CCTV 장비들 중에서 사람 추적기능을 가능 제품은 가격이 대부분 고가이다. 또한 야간에서는 사람의 감지가 힘들다. 본 연구에서는 키넥트의 골격 추적기능과 음성인식 기능을 활용하여 실시간 영상 녹화 프로그램을 개발하고자 하며, 개발된 프로그램은 키넥트 센서로 영상을 실시간 녹화하고 침입자에 대한 움직임을 자동 추적하여 녹화하는 DVR 시스템을 제안하고자 한다. 또한 야간에서는 깊이(Depth) 영상을 이용하여 인물을 인식과 추적을 한다. 궁극적으로 키넥트 센서(Kinect Sensor)의 CCTV기능에 대한 활용성을 연구하는데 목적을 가진다.

• Journal of Institute of Control, Robotics and Systems
• /
• v.22 no.9
• /
• pp.745-752
• /
• 2016

We suggest an efficient Simultaneous Localization and 3D Polygon Map Building (SLAM) method with Kinect depth sensor for mobile robots in indoor environments. In this method, Kinect depth data is separated into row planes so that scan line segments are on each row plane. After grouping all scan line segments from all row planes into line groups, a set of 3D Scan polygons are fitted from each line group. A map matching algorithm then figures out pairs of scan polygons and existing map polygons in 3D, and localization is performed to record correct pose of the mobile robot. For 3D map-building, each 3D map polygon is created or updated by merging each matched 3D scan polygon, which considers scan and map edges efficiently. The validity of the proposed 3D SLAM algorithm is revealed via experiments.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2011.11a
• /
• pp.240-241
• /
• 2011

자연스러운 3D 실감영상을 감상하기 위해서는 많은 시점의 영상이 필요하며 과거 스테레오 디스플레이 장치로부터 최근 그 시점 수가 크게 늘어난 디스플레이 장치로 기술 발전이 이뤄지고 있으며 이에 따라 다시점 콘텐츠를 생성하기 위한 다양한 기술이 개발되어 있다. 다시점 콘텐츠를 생성하기 위하여 ToF 카메라 및 적외선 패턴을 이용한 방법이 주로 이용되고 있으며 이를 활용한 다시점 콘텐츠 생성을 하는 시도가 이뤄지고 있다. ToF 카메라는 PMD사의 제품 및 SwissRanger 사의 제품이 대표적이며 적외선 패턴을 이용한 방식은 MS사의 Kinect가 대표적이며 본 제품들을 활용한 기술 비교를 통하여 다시점 콘텐츠 생성의 결과 및 이를 비교한 장단점을 구분하였다. PMD사의 ToF 카메라는 두 개 이상의 광원을 사용하여 Depth 추출시에 Hole 영역의 크기가 작으나 ToF 영상의 해상도가 매우 작아 고화질의 콘텐츠를 생성하기 위하여 별도의 영상처리 알고리즘이 요구되었다. 반면 MS사의 Kinect는 Depth 영상의 해상도가 상대적으로 커서 영상처리 알고리즘의 복잡도가 작아지나 Depth 추출을 위한 카메라와 RGB 카메라의 위치가 공간적으로 떨어져 있어 이를 보정하기 위한 알고리즘이 요구되며 다시점 변환시 화질에 있어 상대적으로 떨어지는 것으로 나타났다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.552-555
• /
• 2018

This study developed a gesture recognition method using Kinect sensor and proposed a fighting action control interface. To extract the pattern features of a gesture, it used a method of extracting them in consideration of a body rate based on the shoulders, rather than of absolute positions. Although the same gesture is made, the positional coordinates of each joint caught by Kinect sensor can be different depending on a length and direction of the arm. Therefore, this study applied principal component analysis in order for gesture modeling and analysis. The method helps to reduce the effects of data errors and bring about dimensional contraction effect. In addition, this study proposed a modified matching algorithm to reduce motion restrictions of gesture recognition system.