Research On Image Reconstruction And Target Detection Based On Linear Array CCD Rotation Scanning

Linear array CCD rotation scanning technology is an effective way to achieve high-speed imaging and target detection.Our research team designed a new linear scanning CCD imaging system for low-altitude imaging.Based on the system,this paper introduced an idea to accomplish the image reconstruction and target detection of the images.In this paper,based on the in-depth data research,a reconstruction model of the real scene image is established based on the pinhole imaging model of the linear CCD scanning system.The influence of system parameters,such as field angle,rotational speed,sampling frequency and effective pixel number of linear array CCD,on the imaging results is analyzed,and the optimal matching condition of the system parameters is obtained,which provides the theoretical basis for the optimization design of the hardware system.Secondly,by using DMA technology of DSP chip and USB2.0 interface technology,the data transmission control module is developed,and the data transmission rate is greater than 5.7MB/s,and solved the problem that the data transmission rate between the Flash memory and the PC.Is too slow.The memory-mapped file technology is used to realize the loading and dynamic display of the scanned sequence images.By using the periodicity of the stationary background in the sequence image,the autocorrelation of the one-dimensional projection of the image is used to solve the fast and accurate calculation of the sampling number per revolution of linear CCD.And the image reconstruction and display module in the visual coordinate system are developed.Based on the image segmentation technique and the moving object detection technique,the region of the target in the reconstructed image is determined,and the target recognition based on the image template matching is carried out on the extracted region,and the target can be identified quickly and accurately.In addition,it is found that the watershed segmentation algorithm has a good segmentation effect on reconstructed images with complex illumination conditions and weak edges.The background subtraction algorithm based on Gaussian mixture model can detect moving objects in complex backgrounds,Template matching can accurately identify targets.