基于机器视觉的人体运动目标智能监控系统设计与研究-安全技术及工程专业论文.docxVIP

北京交通大学硕士学位论文 北京交通大学硕士学位论文 Abstract A bstract Motion object detection and tracking system combined a lot of advanced sub- jects’technologies such as image processing，pattern recognition，artificial inteUi- gence，autocontrol and computer science．The visual surveillance was broadly ap- plied in robot vision，visual pre-warning，machine navigation，traffic manage，mul- timedia teaching，human-machine interaction and important location safe monitor． With more and more people take attention to safety,safe monitor became very important．Detection and auto-tracking motion people at the complicated environ— ment Was studied in this paper，and which made motion image at the video watch area．This paper contains mainly two parts：First，detecting the motion object po- sition by image processing；second，tracking motion object by motion step motor， which made the motion object always at the video monitor area． At first，the status of research and development on motion object detection and tracking Was studied and analyzed．By comparing many video monitor systems， a novel PT(Pan—Tilt)tracking system was designed and completed in this paper． The system hardware includes step motor and motor driver L297／L298．controller which based on P89V51 single chip computer system and RS．．232 serial port com-- munication．And the connection and communication were realized by MODBUS protoc01． Image processing was the key part for detecting the object and its position． First，image acquisition and image segment such as median filtering，binary image， mathematical morphology and edge detection were studied and completed．A novel edge detection function Was used and it gains the more cIearler and bigger image edge than the Canny operator．Second，image recognize and tracking，for the real- time needs，a new fast model matching way Was presented by decreasing the search matching area，and the popular Mean Shift algorithm Was also a