Sensitivity Evaluation of Radiographs Based on Multi-dimensional Data Coupling Analysis

Abstract

The sensitivity evaluation of image quality indicator (IQI) is an important factor to judge the qualification of radiographs. In order to solve the problems of subjective and low efficiency of manual evaluation method, as well as the difficulty in IQI features extraction and accurate evaluation of sensitivity, a method for determining the sensitivity of radiographs based on coupling analysis of two-dimensional image and one-dimensional gray data is proposed. Firstly, the IQI location technology based on gradient region growth is proposed to realize IQI location in complex background in digital image. Secondly, the IQI image detail enhancement and noise reduction are realized through angle self-correction and enhancement processing. Thirdly, the two-dimensional image data is transformed into one-dimensional gray change data by integral projection method, and a method of IQI feature recognition and sensitivity evaluation based on gray data extreme value change analysis is proposed. The proposed method is evaluated using the actual industrial radiographs. The evaluation results demonstrate the effectiveness and superiority of the proposed method in automatic sensitivity evaluation of radiographs under complex background and low contrast.