Research On Sub-pixel Segmentation Technology Of Image And Its Application In Aircraft Outer Contour Segmentation

Aircraft outer contour recognition plays an important role in the aviation field.Due to the poor accuracy of the outer contour of the image obtained by the traditional image segmentation method,the accuracy rate is reduced in the later recognition process.In order to solve this problem,this paper studies the sub-pixel segmentation technology,and optimizes the segmentation effect of multiple images,and finally applies it to the segmentation process of the outer contour of the aircraft.Combined with the current sub-pixel segmentation algorithm,it is difficult to achieve the accuracy of the aircraft image processing.This paper studies the problems existing in the sub-pixel segmentation process of aircraft images,and uses a series of improved algorithms to improve the accuracy and noise immunity of the segmented image.Ability and improve the computational efficiency of the whole process,and apply the algorithm studied in this paper to the segmentation process of the outer contour.The main research work of this paper is as follows:(1)In the process of sub-pixel positioning of images,the subjective and objective methods are used to judge the pros and cons of each positioning algorithm.It is found that the commonly used interpolation positioning method is difficult to solve the problem of inconsistent interpolation efficiency and computational efficiency.Therefore,this paper improves a positioning method based on hierarchical interpolation,which is processed according to gray information and then interpolated for each layer.The method can further improve the image positioning accuracy after interpolation,and ensure the calculation efficiency can meet the operation requirements of processing images in the aviation field.(2)In the past,the image segmentation method based on sub-pixels only processed the individual information of sub-pixels,but neglected the relationship between sub-pixel neighborhoods.In image segmentation,it is often interfered by some noise,background and other factors.The process may cause the image model to fall into a locally optimal state when optimized.This paper introduces Markov random field and Bayesian estimation(MRF-MAP)to apply statistical methods to sub-pixel information in images,and establish a functional model of sub-pixels in images,and then use cross-visual cortical model(ICM)optimization.The theory is optimized.In the optimization process,a clustering method is improved,and the mean point drift clustering is used to realize the initial point marking,which can effectively reduce the background information interference to the target and has strong anti-noise ability.(3)For the image segmentation method,the traditional segmentation method may have the problem of under-segmentation or over-segmentation.This paper uses the genetic neural network method to achieve accurate segmentation of the target image and solve the problem of under-segmentation or over-segmentation.In addition,in view of the problem of edge inhomogeneity after image segmentation in genetic neural network method,this paper combines fuzzy theory with genetic neural network,and uses the membership function to degrade the membership degree in the edge detection process to complete the image segmentation process.The image segmented by the genetic fuzzy neural network method is better,and the obtained image segmentation accuracy can reach 0.02 or more.(4)Applying the relevant algorithm studied in this paper to the segmentation process of the outer contour of the aircraft,the effectiveness of the proposed algorithm is verified,and the accuracy in the later identification and judgment process can be guaranteed.