Research On High-speed Dynamic Image Processing Algorithm And Its Applications

In this paper, we designed a visual system to complete the high-speed imageacquisition and processing. In order to achieve the requirements we acquire images ofthe moving target by predict it’s trajectory, which means we only acquire the region thatmoving target might arise base on the result of the trajectory prediction, through thisway we reduce the area of image we should process to improve the speed of imageacquisition and processing. We completed the integral design of the visual system, andtested the performance of the system through experiment.Firstly, connected the camera, NI-CVS and the host computer to build a real-timevisual system, based on LabVIEW and NI Real-Time Operation System we create amulti-threaded applications to complete image acquisition task and the networkcommunication task. With the real-time visual system we can do the further research.The following is the research on image processing algorithm, which mainlyincludes image format transformation, image segmentation, image filtering and get thecenter of gravity. We compared the advantage and disadvantage of the commonly usedalgorithms, and finally identify the most suitable one for our application.At last, we used the Extended Kalman Filtering and three spline interpolation toprediction the trajectory of the moving target. Combined with the real-time visualsystem and image processing algorithm, we design the whole visual system, and testedthe performance of the system through experiment.