Strategy Research On The Detection And Tracking Of Moving Objects Based On SIFT

Detection and tracking of moving objects is always a challenging and importantresearch topic in the computer vision field. The subject combines with the frontierknowledge of physics, applied mathematics, computer science and many other disciplines.In addition it’s successfully used in industrial inspection, intelligent transportation,medical diagnostics and other areas. So far, based on the difference environment ofmoving objects, some research is focused on the representation goals, feature selectionand modeling were done in this subject, and the presence of the moving objectenvironment can be divided into static background and dynamic background. Based on adynamic context in this paper, taking a two degrees of freedom pan-tilt as the controlledobjects, using image processing methods to study the detection and tracking of movingobjects, and designing a real-time control systems for the pan-tilt. The main work is asfollows:Firstly, analyzing the structure and working principle of two degrees of freedompan-tilt by reading and analyzing a large number of relevant literatures. Some researchstatus about moving object detection and tracking algorithms were realized, and theadvantages and disadvantages of these algorithms were summarized by compiling existingalgorithms and experimental,.Secondly, an improved solution was proposed for the existing motion object detectionalgorithm. For solving the problem of the moving object detection when the camera ismoving, this paper proposed a method for moving object detection that combined the SIFT(Scale Invariant Feature Transform Scale Invariant Feature Transform) image blockfeature matching algorithm with the local differential. Through the experiment that: themethod reduced the calculated area of the image, speeded up the data processing, andenhanced the real-time characteristics of the system.Again, an improved solution was proposed to the existing moving object trackingalgorithm. For solving the problem of the moving object tracking in dynamic environment,an object tracking algorithms was proposed based on SIFT, and the pedestrian tracking experimental results demonstrate the effectiveness of the program.Finally, two degrees of freedom electric pan-tilt was set up by using servos. Thestructure and parameters of servos and servo pan-tilt were analyzed, the motion controlstrategy was proposed based on the image processing, and achieving a real-time control ofthe pan-tilt.