Research On High-resolution Ultrasinic Detection Technology

Ultrasound occupies an important position in medical diagnosis and industrial detection with its unique advantages such as low cost,no radiation,strong penetrating power,and high directivity.Different media have different acoustic impedance.When ultrasonic waves encounter impurities or interfaces,they will have obvious reflection and scattering phenomena.We construct images according to the reflection intensity to distinguish the positions of different flaws or tissues.With the in-depth study of ultrasound imaging technology,researchers have found that ultrasound imaging technology still has many problems.For example,the internal structure of the inspected material and noise in industrial inspections will affect the echo of target,and the attenuation speed will increase as the depth increases.When two targets or the interval between layers is small,the echoes will overlap,which will cause deviations in the estimated results and affect the quality of imaging.The traditional line scan imaging mode reduces the imaging frame frequency due to multiple emission and focusing,which cannot meet the current medical and industrial imaging requirements for moving targets.However,the plane wave is currently the most widely used imaging technology to increase the frame rate.But the quality of the image is poor due to its unfocused emission.Coherent plane wave compounding(CPWC)is a major improvement after plane waves.Although the image quality can be improved by superimposing plane waves in different directions,due to the strong coherence between plane waves,imaging resolution still needs to be improved.In view of the above existing problems,this paper mainly studies the following two aspects:(1)In this paper,a parameter estimation algorithm of ultrasonic echo signal through joint sparse representation and flower pollination algorithm is proposed,and the performance of the algorithm is illustrated by using Gaussian model and CramerRao Lower Bounds(CRLB).According to the principle of A-scan imaging,the envelope of Gaussian signal contains the key factor of ultrasonic nondestructive testing-time of arrival(TOA).However,due to the interference of noise,the extracted envelope of signal is often incomplete,which will affect the final estimation result.Therefore,in this paper an improved matching pursuit algorithm(IMP)is first used to sparse the echo signal to filter out noise,then Hilbert transformation method is used to extract the echo envelope,and finally uses the flower pollination algorithm to estimate the parameters contained in the echo signal.Matching pursuit algorithm(MP)is a common method in the field of sparse representation.Aiming at its characteristics of greed,this paper proposes an improved MP method of sub-search small-step superposition.Experiments show that the algorithm improves the convergence speed and accuracy of the MP algorithm.Flower pollination algorithm has a strong ability to estimate multiple objective functions.The combination of the two can accurately estimate the parameters of the multi-echo signal and reflect the relationship among the echoes.(2)A-scan imaging is mainly used to judge the location of the flaws along the ultrasonic propagation line.However,if we want to observe the information of a certain cross section inside the material,the ultrasonic phased array imaging method is usually used.At present,the imaging algorithm with the highest frame rate in ultrasonic phased array is plane wave imaging,which has a high frame rate,but due to its unfocused characteristics,the beams have strong coherence and poor imaging quality.Aiming at the characteristics,an adaptive imaging method for ultrasound coherent plane wave compounding based on polar coherence factor(PCF)is proposed in this paper,which uses the imaging point as a weighted object.And generate PCF weighting factors by analyzing the polarities of imaging points in different plane waves.According to the polar performance of imaging points in different plane waves,side lobes and noise interference can be well suppressed to improve imaging resolution and contrast.This has also been verified on the data set provided by the Plane Wave Imaging Challenge in Medical Ultrasound.