Panoramic Image Preprocessing Based On FPGA

Panoramic vision pretreatment to become a research hotspot in recent years because ofits wide range of value. In the treatment system,the accuracy of the hardware design andsoftware design flexible to become one of the current urgent problems. In practicalapplications, the concern of the range of perspective and scale of the image is relatively high.Carried out studies on this issue.Collection methods and image formats from the system lensto start the analysis.Panoramic visual sensor in part by the hyperboloidal mirror, two CCDcamera, high-resolution camera composition.The image preprocessing system based on thispanoramic vision sensors.The panoramic image pre-processing platform consists ofFPGA-based core board and peripheral board.FPGA’s features is flexible programming, andparallel processing capabilities is strong. The core board is composed of five parts: FPGAminimum system, memory chips (SDRAM, Flash and SRAM), configurationcircuit,extended interface and power supply circuit. The core of the system board isindependent, for different functional requirements can be configured in different peripheralboards to achieve. The design of the peripheral board includes the following parts: CameraLink image acquisition circuit, display circuit, communication circuit, a storage circuit,man-machine communication circuit and power supply. Image capture and display processusing ping pong operation.System image processing part comprises a panoramic image stitching and expansion.Image stitching part mainly adopts three kinds of new stitching algorithm, which is based onthe central location of the image mosaic algorithm, traversing the registration imagemosaicing algorithm and particle swarm based on image mosaic algorithm. The experiment ofsimulation, based on the central location of the targeted image stitching algorithm, is suitablefor the circular feature image. Traverse the registration image mosaicing algorithm isaccording to the features of the given.Image Stitching Based on Particle Swarm algorithm isbased on particle swarm optimization, through the improvement of the algorithm to solve thedefects into a part of the optimal. Finally, through debugging and software simulation, implementation of panoramic imagecollection and display LCD screen expansion. Stitching algorithm for the simulation effect isgood.