基于VR技术的X射线图像安检危险品自动识别
    点此下载全文
引用本文:田 萌.基于VR技术的X射线图像安检危险品自动识别[J].计算技术与自动化,2022,(1):123-128
摘要点击次数: 230
全文下载次数: 0
作者单位
田 萌 (中国海关管理干部学院河北 秦皇岛 066000) 
中文摘要:针对当前X射线图像安检危险品识别方法未采集模糊静态图像目标,导致安检危险品图像呈现效果较差、危险品识别率较低、识别时间较长的问题,提出了基于VR技术的X射线图像安检危险品自动识别方法。通过X射线获取安检危险品成像,采用VR技术采集模糊静态图像目标,利用光学成像原理分层处理模糊静态图像目标,获取模糊静态图像目标亮度层和细节层,压缩模糊静态图像目标自适应分区,实现危险品图像目标重现。基于沃尔什变换方法提取危险品图像纹理特征,构建BP神经网络模型,反复调整权值和阈值并进行训练,保证输出误差最小化,实现X射线图像安检危险品自动识别。实验结果表明:所提方法的安检危险品图像呈现效果较好,能够有效提高危险品识别率,缩短危险品识别时间,具备良好的危险品识别性能。
中文关键词:VR技术  X射线图像  安检危险品  自动识别  图像目标重现  BP神经网络
 
Automatic Identification of Dangerous Goods in X-ray Image Security Inspection Based on VR Technology
Abstract:In view of the current X-ray image security inspection method for dangerous goods identification, which does not collect fuzzy static image targets, resulting in poor security inspection of dangerous goods images, low recognition rate of dangerous goods, and long recognition time, an automatic identification method of dangerous good in X-ray image security inspection based on VR technology is proposed. Obtaining security inspection dangerous goods imaging through X-rays, using VR technology to acquire fuzzy static image targets, using optical imaging principles to process fuzzy static image targets in layers, acquiring fuzzy static image targets' brightness layer and detail layer, and compressing fuzzy static image targets for adaptive partitioning, realizing the target reproduction of dangerous goods images. Extracting the texture features of dangerous goods images based on the Walsh transform method, building a BP neural network model, repeatedly adjusting the weights and thresholds and perform training to ensure that the output error is minimized, and realize the X-ray image security inspection of dangerous goods auto recognition. The experimental results show that the proposed method has a better presentation effect of dangerous goods images for security inspection, can effectively improve the recognition rate of dangerous goods, shorten the time of dangerous goods recognition, and have good dangerous goods recognition performance.
keywords:VR technology  X-ray image  security check for dangerous goods  automatic identification  image target reproduction  BP neural network
查看全文   查看/发表评论   下载pdf阅读器