神经网络在岩石多角度偏振光谱识别中的应用  

The application of neural network in identifing multiangular polarized spectrum of rock

作　　者：赵云升[1] 黄睿[1] 胡新礼[1] 张洪波[1] 

机构地区：[1]东北师范大学城市与环境科学学院,吉林长春130024

出　　处：《东北师大学报（自然科学版）》2005年第1期113-117,共5页Journal of Northeast Normal University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(4977105)

摘　　要：　依据岩石多角度偏振光谱反射特征数据,运用神经网络方法,对反射特征相近的岩石加以识别.在识别中以不同方位角、反射角及入射角的反射比等参数作为训练样本及测试样本,训练样本经过网络学习、训练得到神经网络模型,测试样本对神经网络模型进行检验,实验证明将神经网络应用于岩石多角度偏振遥感技术是岩石识别的一种切实可行的方法.By the data of polarized light reflectance feature of rock,with the method of neural network,the work has been done to distinguish from the resemble polarization of rock.During identifying,differentpolarized reflectance of different incidence angle,view angle,azimuth angle were regarded as the training sample,then the model of neural network can be given by network studying and network training of training sample,and the testing sample checkout the model of neural network.The examination prove it is a feasible method to identify rock using neural network on modern romote sensing of polarized light reflectant feature of rock.

关 键 词：神经网络 岩石 偏振反射比 识别 

分 类 号：TP7[自动化与计算机技术—检测技术与自动化装置]