基于边缘计算实现三维勘测数据的自动分类方法研究

Research on Automatic Classification of 3D Survey Data Based on Edge Computing

三维勘测数据的自动分类效果,直接影响数据的利用程度和应用效果,因此,提出基于边缘计算实现三维勘测数据的自动分类方法。采用边缘计算预处理三维勘测数据,在边缘设备中嵌入式中心库,建立边缘计算信任模型,保证数据安全性;采用广义幂变换Zscore算法标准化处理预处理后的三维勘测数据后,将其输入基于多尺度卷积神经网络中进行训练建立三维勘测数据的自动分类模型。测试结果表明:该方法能够保证三维勘测数据并发访问和调用安全,自动分类性能良好,最高损失值为0.18,分类结果的汉明损失指标值均在0.025以下,可以满足实际应用需求。

The automatic classification effect of 3D survey data directly affects the utilization degree and application effect of data.There-fore,an automatic classification method of 3D survey data based on edge calculation is proposed.The edge computing is used to preprocess the 3D survey data,and the central library is embedded in the edge device to establish the edge computing trust model to ensure the data security.After the preprocessed 3D survey data are standardized by the generalized power transform zscore al-gorithm,they are input into the multiscale convolution neural network for training to establish an automatic classification model of 3D survey data.The test results show that this method can ensure the security of concurrent access and call of 3D survey data,and the automatic classification performance is good.The maximum loss value is 0.18,and the Hamming loss index values of classification results are all below 0.025,which can meet the needs of practical application.