Study On Data Acquisition And Scatter Correction System Of Dual Energy X-ray Absorptionmetry

Dual-energy X-ray absorptiometry has the advantages of high accuracy and low dose in bone density measurement and is widely used in the diagnosis and treatment of clinical osteoporosis.The dual-energy Xray bone densitometers currently used in clinics still have disadvantages such as slow scanning speeds.In order to further increase the scanning speed and reduce the absorbed dose of patients,a data acquisition system for dual-energy X-ray bone densitometry was designed and studied.The problem of correcting ray scattering and the method for measuring bone mineral density were discussed.The main research work included:First,the bone mineral density measurement method,the bone density value calculation method and the development status of the dual-energy X-ray bone densitometer were investigated.According to the different parts of the measurement and scanning methods,they were classified,and the development trend of dualenergy X-ray bone densitometer was prospected.Second,the system structure and principle of dual-energy X-ray bone densitometer are introduced.Cone-beam DEXA data acquisition system was designed.Data acquisition system includes X-ray imaging and embedded control.In the imaging section,a k-edge filter method was proposed as a dual-energy ray generation scheme,and an amorphous silicon flat panel detector was used for imaging design.A twodimensional stepped phantom is designed.The phantom contains a combination of aluminum and plexiglass with different thicknesses,thus improving the efficiency of data collection.In the embedded control part,Exynos4412 is used as the control system microprocessor,Ethernet is used to communicate with the host computer,and a scheme for controlling a stepping motor using an FPGA chip is proposed.Third,the cone beam ray scattering correction method was studied.A scattering kernel deconvolution algorithm is proposed.The algorithm first performs Wiener filtering on the image to reduce noise,then classifies the signal of each pixel in the graph,finds the corresponding scattering kernel function for deconvolution,and finally superimposes all The deconvolution result of the category.In order to verify the effectiveness of the algorithm,simulation data was obtained using the EGSnrc software based on the Monte Carlo method.The simulation data was brought into the algorithm and the results showed that the scattering noise was suppressed.Fourth,the calculation method of bone mineral density in dual-energy X-ray bone densometer was explored.The process of calculating the bone density value is introduced.First,the region of interest is segmented from the image,then the corresponding bone density value is found according to the gray value of each point,and finally the average bone density within the region is calculated.In order to segment the region of interest accurately,six classic image segmentation algorithms are compared.The Canny edge detection algorithm and the smoothed threshold segmentation algorithm have the best effect.In order to find the relationship between high and low energy gray values and bone density,two-dimensional experiments were performed using two-dimensional stepped phantoms to obtain high and low energy projections of the phantoms and accurate bone density values.Then the two polynomial surface fittings were used to find both.The relationship between the verified average error is within 5%,which proves that the method is feasible.