针对单幅图像中的行人检测问题,提出了基于自适应增强算法(Adaboost)和支持向量机(Support vector machine,SVM)的两级检测方法,应用粗细结合的思想有效提高检测的精度.粗级行人检测器通过提取四方向特征(Four direction features,FDF)...针对单幅图像中的行人检测问题,提出了基于自适应增强算法(Adaboost)和支持向量机(Support vector machine,SVM)的两级检测方法,应用粗细结合的思想有效提高检测的精度.粗级行人检测器通过提取四方向特征(Four direction features,FDF)和GAB(Gentle Adaboost)级联训练得到,精密级行人检测器用熵梯度直方图(Entropy-histograms of oriented gradients,EHOG)作为特征,通过支持向量机学习得到.本文提出的EHOG特征考虑到熵,通过分布的混乱程度描述,具有分辨行人和类似人的物体能力.实验结果表明,本文提出的EHOG、粗细结合的两级检测方法能准确地检测出复杂背景下不同姿势的直立行人,检测精度优于以往Adaboost方法.展开更多
This study examines the seasonal connections between the interannual variations in LOD （length of day）/ AAMglobe （the relative atmospheric angular momentum for the whole globe） and the ENSO-like SST （El Nifio/ So...This study examines the seasonal connections between the interannual variations in LOD （length of day）/ AAMglobe （the relative atmospheric angular momentum for the whole globe） and the ENSO-like SST （El Nifio/ Southern Oscillation-like sea surface temperature） pattern and corresponding zonal and vertical circulations. Consistent with previous studies, the ENSO-like SST impact the following season LOD/AAMglobe, with the strongest correlations in DJF （December, January, and February）, when it is likely to be the peak E1 Nino/La Nifia period. Lag correlations between the interannual variations in LOD/AAMglobe and surface temperature, and the interannual variations in LOD and both zonal circulation and vertical airflow around the equator, consistently indicate that the LOD/AAMglobe reflect the potential impacts of variations in the Earth＇s rotation rate on the following season＇s sea surface temperatures （SST） over the tropical central and eastern pattern is located）. Pacific （where the ENSO-like SST Moreover, the centers of strongest variation in the AAMcolumn （the relative atmospheric angular momentum for an air column and the unit mass over a square meter） are located over the mid-latitudinal North Pacific in DJF and MAM （March, April, and May）, and over the mid-latitudinal South Pacific in JJA （June, July, and August） and SON （September, October, and November）. This suggests that the AAMcolumn over the mid-latitudinal Pacific around 30°N （30~S） dominate the modulation of Earth＇s rotation rate, and then impact the variations in LOD during DJF and MAM （JJA and SON）.展开更多
文摘针对单幅图像中的行人检测问题,提出了基于自适应增强算法(Adaboost)和支持向量机(Support vector machine,SVM)的两级检测方法,应用粗细结合的思想有效提高检测的精度.粗级行人检测器通过提取四方向特征(Four direction features,FDF)和GAB(Gentle Adaboost)级联训练得到,精密级行人检测器用熵梯度直方图(Entropy-histograms of oriented gradients,EHOG)作为特征,通过支持向量机学习得到.本文提出的EHOG特征考虑到熵,通过分布的混乱程度描述,具有分辨行人和类似人的物体能力.实验结果表明,本文提出的EHOG、粗细结合的两级检测方法能准确地检测出复杂背景下不同姿势的直立行人,检测精度优于以往Adaboost方法.
文摘This study examines the seasonal connections between the interannual variations in LOD （length of day）/ AAMglobe （the relative atmospheric angular momentum for the whole globe） and the ENSO-like SST （El Nifio/ Southern Oscillation-like sea surface temperature） pattern and corresponding zonal and vertical circulations. Consistent with previous studies, the ENSO-like SST impact the following season LOD/AAMglobe, with the strongest correlations in DJF （December, January, and February）, when it is likely to be the peak E1 Nino/La Nifia period. Lag correlations between the interannual variations in LOD/AAMglobe and surface temperature, and the interannual variations in LOD and both zonal circulation and vertical airflow around the equator, consistently indicate that the LOD/AAMglobe reflect the potential impacts of variations in the Earth＇s rotation rate on the following season＇s sea surface temperatures （SST） over the tropical central and eastern pattern is located）. Pacific （where the ENSO-like SST Moreover, the centers of strongest variation in the AAMcolumn （the relative atmospheric angular momentum for an air column and the unit mass over a square meter） are located over the mid-latitudinal North Pacific in DJF and MAM （March, April, and May）, and over the mid-latitudinal South Pacific in JJA （June, July, and August） and SON （September, October, and November）. This suggests that the AAMcolumn over the mid-latitudinal Pacific around 30°N （30~S） dominate the modulation of Earth＇s rotation rate, and then impact the variations in LOD during DJF and MAM （JJA and SON）.
