Abstract:
PURPOSE: A pedestrian detection method of a pedestrian detection device is provided to perform pedestrian detection on a search window using a second classifier with high accuracy after reducing the number of search windows, thereby performing highly accurate detection of an object while reducing the complexity of detection procedure and power consumption. CONSTITUTION: A pedestrian detection device obtains an image from a digital image device and performs blocking of search windows(210,220). The pedestrian detection device selects a specific block from blocks determined by a pre-learned classifier(230) and produces a specific vector of HOG features from the selected block(240,250). The pedestrian detection device calculates a SVM(Support Vector Machine) response value using the produced feature vector, and performs a first object detection by applying the response value to an AdaBoost Classifier(260,270). If a pedestrian is detected at the first object detection, the pedestrian detection device performs a second object detection to the search window(280). [Reference numerals] (210) Obtain an image; (220) Performs blocking of search windows; (230) Pre-learned classifier; (240) Select a specific block; (250) Produce a specific vector; (260) Calculates a SVM response value; (270) Perform a first object detection; (280) Perform a second object detection; (290) Output decision; (AA) No; (BB) Yes
Abstract:
전방위 카메라 및 팬/틸트/줌 카메라를 이용한 무인 감시 장치 및 방법이 개시된다. 제1영상획득부는 감시대상 영역 전체에 대해 전방위 카메라에 의하여 촬영된 복수의 전체영상프레임을 일정한 시간 간격으로 획득한다. 객체추출부는 복수의 전체영상프레임 중 현재 영상프레임을 구성하는 화소들과 현재 영상프레임에 시간적으로 앞서는 이전 영상프레임을 구성하는 화소들의 화소값의 차를 기초로 현재 영상프레임으로부터 객체영상을 추출한다. 제2영상획득부는 감시대상 영역 중에서 객체영상에 대응하는 감시대상 객체에 대해 팬/틸트/줌 카메라에 의하여 촬영된 객체영상프레임을 획득한다. 제어부는 전체영상프레임 내에서 객체영상의 위치를 기초로 감시대상 객체의 움직임을 추적하여 팬/틸트/줌 카메라의 회전 및 주밍(zooming) 동작을 제어한다. 본 발명에 따르면, 감시대상 영역 전체에 대해 전방위 카메라에 의하여 촬영된 전체영상프레임 및 감시대상 영역에 진입한 감시대상 객체에 대해 팬/틸트/줌 카메라에 의하여 촬영된 객체영상프레임을 모두 사용함으로써, 복수의 무단침입 객체가 존재하는 경우에도 이를 모두 감지할 수 있으며, 무단침입 객체의 상세정보도 얻을 수 있다. 무인 감시, 전방위 카메라, 팬/틸트/줌 카메라
Abstract:
PURPOSE: An unmanned surveillance apparatus using an omni-directional camera and a pan/tilt/zoom camera and a method thereof are provided to assign each intrinsic identification code to object images to continuously photograph the same surveillance target object, thereby increasing performance of finding detailed information of an object. CONSTITUTION: The first image obtaining unit(110) obtains all image frames photographed by an omni-directional camera for the entire surveillance target area at a fixed time interval. An object extracting unit(120) extracts an object image from a current image frame based on differences among values of pixels configuring a previous image frame. The second image obtaining unit(130) obtains an object image frame photographed by a pan/tilt/zoom camera for a surveillance target object corresponding to the object image from the surveillance target area. A controller(140) controls rotation and zooming operations of a pan/tilt/zoom camera by tracking movement of the surveillance target object based on a location of the object image in the entire image frame.
Abstract:
The present invention provides an augmented reality imaging-based sightseeing guide apparatus which easily matches the augmented reality image visualized on the reality of outdoor environment. An embodiment of the apparatus includes a camera part outputting video data corresponding to the reality of an arbitrary sightseeing place; a display part displaying the augmented reality image visualized on the reality of the sightseeing place; and a controller detecting a natural marker in the reality of the sightseeing place on the basis of the image data and controlling the display part to visualize the augmented reality image at the location of the natural marker.