递变能量最佳成像管电压的幂指数预测模型研究

Optimum Tube Potential Prediction of Power Exponent Model on Variable Energy Imaging

摘要:

目前变能量成像管电压选择方法中,没有根据检测物体已采回图像信息,智能、快速选择下一帧成像电压的方法；因此,文章提出变能量最佳成像管电压预测算法；该方法通过变能量预扫描,分析递变能量图像序列中有效厚度(高质量区域)、临近厚度(低质量区域)与电压的匹配关系,并基于幂指数拟合建立有效厚度和临近厚度的灰度变换模型,以及临近厚度最佳成像时的能量预测模型,通过模型求解,实现了能量的自适应预测；最后文章以不同厚度钢块为对象,采集递变能量下的图像序列,利用论文算法逐一预测各个厚度最佳成像管电压,并与实际值对比；结果显示在低能时可跨3~4 mm准确预测,高能时可跨7~10 mm预测,精度可以达到95%以上。

Abstract:

Recently, tube potential selection methods of variable energy imaging have not taken the acquired images’ information into consideration, and can not select the potential intelligently and quickly. This paper proposes an Optimum Tube Potential Prediction Method based on variable energy imaging. The method has four steps. First of all, scan the object to obtain image sequence of variable potential. Secondly, find the effective thickness(ET)(high quality area) and near the effective thickness (NET)(low quality area), and the relationship between them and tube potential. Thirdly, build a function model of potential and image gray difference between ET and NET, thus get the optimal potential prediction model of NET. Finally, search the optimal potential for NET based on the two models. At the end of paper, collect graded voltage image sequence of steel blocks with different thickness, then use the method to predict optimal potential for each thickness and compare them with the actual one. Experimental results showed that this method can accurately predict 3 or 4mm at low potential and 7 or 10mm at high potential. Prediction accuracy can reach over 95%.