The Research Of Invariant Moments Based On Radon Transform And PHTs

In Pattern Recognition,compared with original image,the identified image has a large extent distortion,such as the changes of gray and size and so on.As a result,the image description should be insensitive to variety of image distortion.Fortunately,invariant moments just have such characteristics with a highly concentrated image features,and have been widely used.Based on Radon transform and Polar Harmonic Transforms(PHTs),including PCET,PCT and PST,This thesis is proposed the PCST transform which definite by PCT,PST as the complex’s real and imaginary parts separately,and constructed three invariant moments(RPCET,RPCST and RI)and their features by their real and imaginary parts.According the experiments of image recognition,they show that these three invariant moments have a high recognition rate and good noise immunity.Especially,RPCET and RPCST,in the case of noise-free,the correct recognition rate can reach 100%;In the case of the SNR of Gaussian noise is zero(strong noise),the rate over 96%.In addition,RI has an obvious advantage of lower computational complexity than RPCET and RPCST.Moreover,this thesis is studied and implemented the airport runway battle damage assessment(BDA),based on fuzzy comprehensive evaluation(FCE).According to the basic component structure of the airport runway,we propose the battle damage assessment system of airport runway,based on geometry,texture and overall characteristics,using the knowledge base reasoning(KBR),and simulates the design of the system.The simulation results show that this arithmetic is feasible and effective in the way of the runway BDA.