Research On Normalized Signal To Noise Ratio About Weld-ray Image Based On Computed Radiography

Computed radiography(CR) technology, and part of its advantages, such like energy conservation, environmental protection, high speed and low cost of detection, has performance a good prospect of application in the industrial non-destructive testing area. And currently the Standard of ISO17636-2-2013 that domestic CR technology referred to, prescripts to measure image qualities by normalized signal to noise ratio(SNRN). Researched on the basis and influencing factors of SNRN,this paper provides the basis of domestic CR testing standard.By changing the X-ray tube voltage, exposure and focus, using of different types of imaging(Image Plate referred IP) board andscanning parameters of laser scanner(scan step, laser power, signal gain) to study the variation of imageresolution(SRb) and conventional signal to noise ratio(SNR), the results are: Appropriate increase of tube voltage, exposure and focus, can improve the image SRb and conventional SNR, but the selection ofhigh-definition IP board, high scan step, lower laser power and signal gain can improve the image SRbbut cause a lower image SNR, the results show the existence of certain problem using a separate image SRb and SNR evaluation detection sensitivity of CR image, because changingthe IP board type and scanning parameters affect the SRbof CR system, high-definition systems will get a better image SRb but the noise level will be higher. Therefore, this paper proposes the use of ISO 17636-2-2013 "normalized signal to noise ratio" concept, the SNR and CR systems SRbare be unified into one index to evaluate the detection sensitivity.In this paper, a special 304 austenitic stainless steel with 59 different depth and shape of the crack mb by welding.Byusing Dandong XXG2005 ray machine, 13 # steel linear IQI, Double wire IQI, Ruike HR IP board and HPX-1 CR scanner, INDUSTREX andImageJ image scanning and processing software, to change the process parameters for testing of welding, getthe SNRN of image according measurement and calculation methods about SNRN in ISO17636-2-2013,results show that: when the SNRN valuesmeasured in the weld heat affected zone greater than 168 can be obtained to ensure the effective detection of 59 cracks. Considering the practical application of measurement and calculation SNRN has some limitations, this paper studies the relationship between the average gray value and SNRNof image, results show that: under the conditions of stationary of IP board type and scanning parameters, if the average gray value of image above the threshold, SNRNcansatisfy the detection sensitivity requirements, this can be used to quickly evaluate the CR image detection sensitivity.