Study On PMMW Image Detecting Technology

With the maturation and development of microwave,MMW(Millimeter Wave),for its advantages of the less disturbance from environment and the availability in all-weather, etceteras,is turning into a very important technique in military,daily life and medical science.The rapid development of MMW technique brings huge convience to human-being.However,in the application,the edges of MMW image sometimes can't be detected clearly and can't tell the boundry between object and background,for its big noises and low resolution.This depresses the significance of MMW image,and limits the application of MMW.With the goal to identify the PMMW image,this passage do some research on the hardware system at the first place.The traditional front-end hardware system for PMMW scanning and data-collecting controling is redundant in componet-using,complicated in structure,too much in power-consumption.To overcome the setbacks,passage brings out he PMMW image system with TMS2812 DPS chip as the controller,including the study on the structure of front-end equipment.Theoretically,this system is able to reduce the power-consumption and simplify the system structure.When it comes to the edge detection of PMMW image,this passage introduce some classic edge-detection operator,after analyzing,Canny operator has better effect.However,the traditional Canny algorithm uses pre-set factors during the process,which is very hard to comfirm the details automatically and difficult to protect edges and eliminate the noises at the same time.this passage puts forward a auto-ajust Canny algorithm based on Kirsch gradient amplitude and threshold confirmation.This algorithm improve the process of Gauss filtering and improve the process of threshold confirmation for binaryzing MMW image, auto-ajust the factors.And substituting the original gradient amplitude for Kirsch operator,reflects the diffiences between the neigbor points and center point,solve the problems of traditional algorithm.However,the effects of edge detection in image with close gray-level points is not good yet.In this way,in the future study,it still requires more efforts.