| 基于改进InceptionV3模型的常见矿物智能识别 |
| |
| 引用本文： | 彭伟航,白林,商世为,唐小洁,张哲远.基于改进InceptionV3模型的常见矿物智能识别[J].地质通报,2019,38(12):2059-2066. |
| |
| 作者姓名： | 彭伟航 白林 商世为 唐小洁 张哲远 |
| |
| 作者单位： | 成都理工大学网络安全学院, 四川 成都 610059;自然资源部地质信息技术重点实验室, 北京 100037,自然资源部地质信息技术重点实验室, 北京 100037;成都理工大学数学地质四川省重点实验室, 四川 成都 610059;成都理工大学管理科学学院, 四川 成都 610059,成都理工大学地球科学学院, 四川 成都 610059,成都理工大学地球科学学院, 四川 成都 610059,成都理工大学地球科学学院, 四川 成都 610059 |
| |
| 基金项目： | 国家重点研发计划《基于地质云的地质灾害基础信息提取与大数据分析挖掘》（编号：2018YFC1505501）和成都理工大学国家级大学生创新创业训练计划项目《基于人工智能方法的岩石识别技术研究》（编号：201810616003） |
| |
| 摘 要： | 以常见的16类矿物作为研究对象,收集每一类矿物约1000张图像,按比例划分为训练集、验证集和测试集,通过图像随机选取增加数据的多样性,建立矿物识别InceptionV3模型,训练7万次在测试集上获得81%的识别正确率。通过对损失函数的改进,引入Center Loss损失函数,训练40万次识别准确率提高到86%。对分类的混淆矩阵分析发现,孔雀石等外观特征明显的矿物识别正确率很高,而闪锌矿等与其他矿物容易混淆导致正确率较低。从特征图分析看出,模型很好地提取了孔雀石的放射状特征,矿物图像特征向量聚集程度很高,也说明了模型的可靠性。
|
| 关 键 词： | 矿物图像 矿物识别 人工智能 深度学习 |
| 收稿时间： | 2019/4/17 0:00:00 |
| 修稿时间： | 2019/7/29 0:00:00 |
Common mineral intelligent recognition based on improved InceptionV3 |
| |
| PENG Weihang,BAI Lin,SHANG Shiwei,TANG Xiaojie and ZHANG Zheyuan.Common mineral intelligent recognition based on improved InceptionV3[J].Geologcal Bulletin OF China,2019,38(12):2059-2066. |
| |
| Authors: | PENG Weihang BAI Lin SHANG Shiwei TANG Xiaojie and ZHANG Zheyuan |
| |
| Institution: | College of Network Security, Chengdu University of Technology, Chengdu 610059, Sichuan, China;Key Laboratory of Geological Information Technology, Ministry of Natural Resources, Beijing 100037, China,Key Laboratory of Geological Information Technology, Ministry of Natural Resources, Beijing 100037, China;Geomathematics Key Laboratory of Sichuan Province, Chengdu University of Technology, Chengdu 610059, Sichuan, China;College of Management Science,, Chengdu 610059, Sichuan, China,College of Earth Sciences, Chengdu University of Technology, Chengdu 610059, Sichuan, China,College of Earth Sciences, Chengdu University of Technology, Chengdu 610059, Sichuan, China and College of Earth Sciences, Chengdu University of Technology, Chengdu 610059, Sichuan, China |
| |
| Abstract: | To study 16 kinds of common minerals, the authors collected 1000 images for each type, and then divided them into training set, validation set and test set. Before putting the images into the model, the authors selected a random area of each image for data augmentation. After training the InceptionV3 model with 70000 steps, the authors obtained an 81% accuracy in the test set. Through improving the loss function and introducing the Center Loss, the authors raised the accuracy to 86% after training 400000 steps. The obfuscation matrix shows that, the recognition accuracies for the minerals with obvious appearance characteristics such as malachite are higher while those for other minerals like sphalerite are less due to the obfuscation with other minerals. The analysis of the feature map shows that the model extracts the radial feature of malachite perfectly, and the feature vector of mineral image aggregate is in a high degree, which also can prove the reliability of the model. |
| |
| Keywords: | mineral image mineral recognition artificial intelligence deep learning |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《地质通报》浏览原始摘要信息 |
| 点击此处可从《地质通报》下载免费的PDF全文 |
|
