Research On The Dangshan Fruit Tree Wormhole Non-destructive Detection Based On Filtering Back-projection Image Reconstruction

Dangshan county is known as "the hometown of fruits".It has about 70%fruit tree resources in Anhui province,including nearly 60,000 pear trees which over one hundred years old.The health of fruit trees and the production of fruits directly affect local farmers.Economic benefits.However,the larvae of the crustacean insects,which are mainly based on the beetle,often drill holes in the fruit trees for feeding.It is difficult to distinguish through the appearance of the fruit trees,causing the growth of the fruit trees to decay and even killing the fruit trees,seriously affecting the economic benefits and work of the local people.The personnel carried out the work of protecting the old trees.Therefore,a tool is needed to perform a Iive scan of the fruit trees in tlie field to determine whether the fruit trees are invaded by pests and diseases.The computed tomography technique has the advantages of being able to clearly present the tomographic image of the object to be examined,and is widely used in the industrial and medical industries.At present,although the domestic application of fault technology to the wood non-destructive testing industry,it has not only developed a mature portable testing equipment at the research stage.In order to better solve the problem of non-destructive testing of fruit tree wormholes,this dissertation explores the development of theoretical netliods related to non-destructive testing instruments for portable fruit trees.The main work of this dissertation is as follows:(1)The influence of different filtering function interpolation methods on filtering back-projection algorithm is explored.To explore different filter function,the influence of interpolation method of reconstruction image quality,based on MATLAB8.4 environment,selected Hanning,Hamming Cosine,S-L,R-L five kinds of typical filtering function,as well as the adjacent interpolation method,linear interpolation method,cubic spline interpolation method,three kinds of commonly used interpolation method respectively in the projection data with noise and no noise under the condition of two kinds of filter back projection algorithm for simulation,and use the normalized mean square distance,the normalized mean absolute distance,peak signal-to-noise ratio of three kinds of image quality evaluation index to the reconstruction result is analyzed.The results show that Hanning filter function is the best filter function for image reconstruction,and linear interpolation is the best interpolation method.(2)The method of wormhole extraction is designed and the accuracy of the branch image reconstruction using the filter back-projection algorithm is analyzed.In order to analyze the error of reconstructed wormholes,Hanning filtering function was first used to process the branch CT projection data containing wormholes,and combined with the linear interpolation method to reconstruct the branch tomographic image.For CT image and the actual stems stems in wormhole characteristic information,the analysis of five kinds of threshold segmentation method comparison,finally choose histogram twin peaks law for branches of image segmentation,image morphology operations on image denoising and feature extraction of images,then obtain stems cross section and the area of the wormhole information,finally respectively calculate the stems and branches in the image of CT images of worm hole area.Through the comparison and analysis between the reconstruction results and the actual image,it is shown that the branches fault can be reconstructed accurately by using the filtering back-projection algorithm,and the proportion of the reconstructed wormholes in the branches is 0.11%compared with the actual error.(3)Design and complete the fruit tree wormhole analysis system.In order to process and analyze the collected CT data quickly and conveniently,this dissertation designs and implements the fruit tree wormhole analysis system based on the full analysis of image reconstruction algorithm and feature extraction method,combined with various functions such as menu function,schedule prompt function,image export function and step-by-step analysis operation.In this dissertation,the system can accurately extract the fruit tree wornhole and obtain the image characteristic parameters,which can meet the needs of nondestructive testing of fruit tree wormhole.Through the exploration of fault reconstruction algorithm,wonnhole feature extraction and the design of fruit tree wormhole analysis system,the possibility of nondestructive testing of fruit tree wormhole was verified,and the theoretical basis was provided for the development of fruit tree-wormhole nondestructive testing equipment.