Automated Target Tracking & Pan-tilt Camera Tutor : 고형화 손채봉 Studied by : 오재도 최재형 이희웅 정종윤 2008 Capstone Project.

Presentation on theme: "Automated Target Tracking & Pan-tilt Camera Tutor : 고형화 손채봉 Studied by : 오재도 최재형 이희웅 정종윤 2008 Capstone Project."— Presentation transcript:

1 Automated Target Tracking & Pan-tilt Camera Tutor : 고형화 손채봉 Studied by : 오재도 최재형 이희웅 정종윤 2008 Capstone Project

2  Manufacture the Pan-tilt Camera(servo hs-300)  Realization Target tracking method using image processing (computer language : C++)  Control the servo… Capturing… Processing… Control…

3 Target Tracking using PTZ camera 2008 Capstone Project Object FeatureExplainationExample Motion Boundary boundary tracking Active Contour Model (Snake) Region Calculate Object’s Direction, Velocity Optical Flow Model Learning Each Model -> Tracking [koller 91] Color Using Color Distribution Mean shift Object Tracking Methods Why? We want to know How computer recognize an object as the ‘Object’ (from only image) And we thought As a first step towards a perceptual user interface, a computer vision color tracking algorithm is developed. the context of other school subjects Engineering mathematics probability & Random processes computer imaging Signal&system High level programming Comunication theory…Etc… How Computer recognize Object??

4 Difference between Our goal and result 1. Appearance – Similar figures 2. Algorithm Sensitivity – OTL it is difficult to discern target from not if there are equal-color object in frame 2008 Capstone Project Video capture Color conversion RGB->HSV Histogram back projection New search window Calculate Moments Mean Shift Tracking flow by CAMShift Transmit the Digital signal to controller (COM Port) -> controller : pulse generating product name : Ispaulder Specifications Camera input Connectivity - USB 1.1 port or higher (2.0 recommended) Other - 256 MB of RAM or higher. -Display adapter capable of 16-bit color depth or higher, 2 MB or higher video memory. Pan-Tilt driver -COM port(RS-232, 466) USB2.0 Tracking software Hard Drive Space 300-700 MB free

5 2008 Capstone Project

6 잘된 점 - 팀원간의 업무분담 및 예정대로의 진행 - 추적방식에 대한 지속적인 연구와 세미나 - 개선 되어야 할 점 다른 추적방식과 연동을 통한 알고리즘 취약점 향상 미진한 점 - 완성 후 미비점 보완 설계를 통해 배운 점 팀원간의 협동의 중요성 영상만을 이용한 추적은 processor 의 처리속도와 memory 기술의 발전이 뒷받침 되어야 할 것 같다. 자기평가 2008 Capstone Project 업그레이드는 계속되어야 한다.

7 2008 Capstone Project 영상처리 Library 조사 관련 알고리즘 구현 장비 interface 조사 디자인 고안 이희웅, 정종윤오재도, 최재형 추적방식 / 알고리즘 에 대한 연구 serial Comunication 학습 / 제어 구현

8 2008 Capstone Project