Research On New Infrared Thermal Imaging System Based On Image Fusion

Image is an important means for humans to acquire and record information,and it has an irreplaceable role in the field of photoelectric information detection.In recent years,multi-band fusion imaging technology represented by infrared and visible light bands can extract the dominant features of two-band images and enhance people's ability to recognize and understand images in the detection field.It has been widely used in industrial production and daily life.After extensively reviewing the literature and research related systems,and fully considering the development difficulty and application scalability,the design uses the embedded FPGA platform plus the PC platform system implementation.Focusing on the image fusion system of infrared and visible light,the main research contents of this paper are as follows:1.The imaging principle of infrared detector and visible light sensor is studied.The materials,processes and basic physical principles involved in the current mainstream uncooled resistance thermal radiation focal plane detectors,CCD sensors and CMOS sensors are discussed.Guided by this research,the Dali DLD384 detector was selected as the infrared detector,and the Sony IMX322 CMOS sensor was the visible light sensor.2.The infrared and visible image fusion algorithms and related theories are studied in depth.The research algorithms mainly include: non-uniformity correction algorithm of infrared image,image enhancement algorithm,pseudo color conversion algorithm,image registration algorithm and infrared and visible image fusion algorithm.Based on the above infrared and visible image fusion theory,considering the fusion effect and algorithm operation efficiency,a regional fusion algorithm which can be implemented quickly is proposed.3.Perform hardware design and program design of image acquisition system.The hardware design includes: DLD384 detector control circuit,IMX322 sensor control circuit,FPGA interactive control circuit,SDRAM control circuit,USB3.0 transmission control chip peripheral circuit.Based on the above design,PCB design and PCB board are drawn.The program design includes: FPGA logic control program design,DLD384 detector and IMX322 sensor control program design,SDRAM read/write programdesign,USB3.0 control program design.4.The design of the front-end image acquisition system,combined with the host PC,to build an experimental system of image fusion.In the upper PC,the algorithm is implemented in the OpenCV library using the C++ language,and the GUI interface is written using the Qt framework.The image data collected at the front end was subjected to non-uniformity correction,image registration and image fusion experiments.Finally,the results were evaluated and analyzed.The research in this paper realizes the acquisition of infrared and visible images,and completes image fusion on the PC side.From the perspective of fusion effect,both subjective observation and objective index calculations show that the information of the merged image is greatly increased compared with the single-band image.From the perspective of algorithm efficiency,under the experimental conditions,the single frame fusion processing time reaches about 0.037 seconds,which can be applied to the video stream processing of infrared and visible light fusion.