目的提出一种结合C/S(Client/Server)架构和BRF(Boosted random ferns)算法的移动增强现实应用方案,以保证图像识别算法对于产品外包装的识别性能。方法 BRF是一种高效、鲁棒的特征匹配算法,但由于手机内存及处理器等硬件条件的制约,不...目的提出一种结合C/S(Client/Server)架构和BRF(Boosted random ferns)算法的移动增强现实应用方案,以保证图像识别算法对于产品外包装的识别性能。方法 BRF是一种高效、鲁棒的特征匹配算法,但由于手机内存及处理器等硬件条件的制约,不能直接适用于手机终端。将C/S模式与BRF算法相结合应用于图像特征匹配,并设计实验测试比较文中方案(CS-BRF)与ORB算法的识别速度和匹配精度。结果实验结果表明,相比ORB算法,CS-BRF在识别速度相近的前提下,具有更为优异的识别精度。结论 CS-BRF能够实时准确识别印刷品图像,良好适用于产品包装移动增强现实系统。展开更多
It is difficulties for the computer simulation method to study radiation regime at large-scale.Simplified coniferous model was investigated in the present study.It makes the computer simulation methods such as L-syste...It is difficulties for the computer simulation method to study radiation regime at large-scale.Simplified coniferous model was investigated in the present study.It makes the computer simulation methods such as L-systems and radiosity-graphics combined method(RGM) more powerful in remote sensing of heterogeneous coniferous forests over a large-scale region.L-systems is applied to render 3-D coniferous forest scenarios,and RGM model was used to calculate BRF(bidirectional reflectance factor) in visible and near-infrared regions.Results in this study show that in most cases both agreed well.Meanwhile at a tree and forest level,the results are also good.展开更多
针对轴向柱塞泵实际故障诊断中采集到的故障类数据远少于正常类数据的情况,为提升故障分类精确率,提出了一种基于平衡随机森林(Balanced Random Forest,BRF)的轴向柱塞泵故障诊断方法。BRF算法是随机森林(Random Forest,RF)的改进算法,...针对轴向柱塞泵实际故障诊断中采集到的故障类数据远少于正常类数据的情况,为提升故障分类精确率,提出了一种基于平衡随机森林(Balanced Random Forest,BRF)的轴向柱塞泵故障诊断方法。BRF算法是随机森林(Random Forest,RF)的改进算法,将欠采样方法与RF结合,强化了RF处理非均衡数据的能力。通过开源的UCI数据集对该算法的性能进行了测试,相较于RF以及合成少数类过采样(Synthetic Minority Over-sampling Technique,SMOTE)与RF的组合算法SMOTE-RF,BRF算法在少数类分类精确率方面有所提升。最后,将BRF算法应用于轴向柱塞泵的故障诊断中。结果表明,在类间数据不均衡的条件下,相较于RF及SMOTE-RF算法,BRF算法能够取得更高的故障分类精确率。展开更多
文摘目的提出一种结合C/S(Client/Server)架构和BRF(Boosted random ferns)算法的移动增强现实应用方案,以保证图像识别算法对于产品外包装的识别性能。方法 BRF是一种高效、鲁棒的特征匹配算法,但由于手机内存及处理器等硬件条件的制约,不能直接适用于手机终端。将C/S模式与BRF算法相结合应用于图像特征匹配,并设计实验测试比较文中方案(CS-BRF)与ORB算法的识别速度和匹配精度。结果实验结果表明,相比ORB算法,CS-BRF在识别速度相近的前提下,具有更为优异的识别精度。结论 CS-BRF能够实时准确识别印刷品图像,良好适用于产品包装移动增强现实系统。
基金the Chinese National Natural Science Foundation Project(40701124)the Chinese Hi-tech Research and Development Program Project(2006AA12Z114)
文摘It is difficulties for the computer simulation method to study radiation regime at large-scale.Simplified coniferous model was investigated in the present study.It makes the computer simulation methods such as L-systems and radiosity-graphics combined method(RGM) more powerful in remote sensing of heterogeneous coniferous forests over a large-scale region.L-systems is applied to render 3-D coniferous forest scenarios,and RGM model was used to calculate BRF(bidirectional reflectance factor) in visible and near-infrared regions.Results in this study show that in most cases both agreed well.Meanwhile at a tree and forest level,the results are also good.
