期刊文献+
共找到2篇文章
< 1 >
每页显示 20 50 100
单子叶作物叶片气孔自动识别与计数技术 被引量:5
1
作者 孙壮壮 姜东 +5 位作者 蔡剑 王笑 周琴 黄梅 戴廷波 曹卫星 《农业工程学报》 EI CAS CSCD 北大核心 2019年第23期170-176,共7页
为实现作物叶片气孔的自动识别与快速计数,该研究采用卷积神经网络中高计算效率的YOLOv3算法,开发了一种全自动气孔识别和计数解决方案。该算法优化了物体检测性能,可准确识别显微图像中的气孔。其中,对指甲油印迹法获得照片的气孔检测... 为实现作物叶片气孔的自动识别与快速计数,该研究采用卷积神经网络中高计算效率的YOLOv3算法,开发了一种全自动气孔识别和计数解决方案。该算法优化了物体检测性能,可准确识别显微图像中的气孔。其中,对指甲油印迹法获得照片的气孔检测精确率、召回率和F1值分别为0.96,0.98和0.97,便携式显微镜拍摄法照片气孔检测精确率、召回率和F1值分别为0.95,0.98和0.96,具有很好的鲁棒性。该算法检测速度快,可实现对30帧/s的视频文件进行快速气孔识别,实现了实时检测。此外,采用拍摄的小麦叶片照片进行训练得到的气孔识别模型,还可同时实现对大麦、水稻和玉米等单子叶作物叶片气孔的识别,其中,大麦的检测精确率、召回率和F1值分别为0.94,0.83和0.88;水稻的检测精确率、召回率和F1值分别为0.89,0.42和0.57;玉米的检测精确率、召回率和F1值分别为0.91、0.76和0.83;显示出模型良好的泛化能力。 展开更多
关键词 卷积神经网络 机器视觉 模型 单子叶作物 气孔识别 计数 深度学习 实时检测
下载PDF
Application of multi-component joint inversion in oil and gas exploration:A case study of reservoir and gas saturation prediction of the Xujiahe formation in the PLN area of the central Sichuan Basin
2
作者 Wang Dong He Zhen-Hua +4 位作者 Wang Xu-Ben Li Le Yang Hai-Tao Fu Zhi-Guo Wang Hong-Yan 《Applied Geophysics》 SCIE CSCD 2020年第5期879-889,905,共12页
Multi-component seismic exploration is an important technique in the utilization of P-waves and converted S-waves for oil and gas exploration.It has unique advantages in the structural imaging of gas zones,reservoir p... Multi-component seismic exploration is an important technique in the utilization of P-waves and converted S-waves for oil and gas exploration.It has unique advantages in the structural imaging of gas zones,reservoir prediction,lithology,and gas-water identifi cation,and the development direction and degree of fractures.Multi-component joint inversion is one of the most important steps in multi-component exploration.In this paper,starting from the basic principle of multi-component joint inversion,the diff erences between the method and single P-wave inversion are introduced.Next,the technique is applied to the PLN area of the Sichuan Basin,and the P-wave impedance,S-wave impedance,and density are obtained based on multi-component joint inversion.Through the velocity and lithology,porosity,and gas saturation fi tting formulas,prediction results are calculated,and the results are analyzed.Finally,multi-component joint inversion and single P-wave inversion are compared in eff ective reservoir prediction.The results show that multi-component joint inversion increases the constraints on the inversion conditions,reduces the multi-solution of a single P-wave inversion,and is more objective and reliable for the identification of reservoirs,effectively improving the accuracy of oil and gas reservoir prediction and development. 展开更多
关键词 multi-component joint inversion lithology identifi cation POROSITY gas saturation reservoir prediction
下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部