Few studies have specifically focused on the validation and spatiotemporal distribution of planetary boundary layer height （PBLH） and relative humidity （RH） data in China. In this analysis, continuous PBLH and sur...Few studies have specifically focused on the validation and spatiotemporal distribution of planetary boundary layer height （PBLH） and relative humidity （RH） data in China. In this analysis, continuous PBLH and surface-level RH data simulated from GEOS-5 between 2004 and 2012, were validated against ground-based observations. Overall, the simulated RH was consistent with the statistical data from meteorological stations, with a correlation coefficient of 0.78 and a slope of 0.9. However, the simulated PBLH was underestimated compared to LIDAR data by a factor of approximately two, which was primarily because of poor simulation in late summer and early autumn. We further examined the spatiotemporal distribution characteristics of two factors in four regions--North China, South China, Northwest China, and the Tibetan Plateau. The results showed that the annual PBLH trends in all regions were fairly moderate but sensitive to solar radiation and precipitation, which explains why the PBLH values were ranked in order from largest to smallest as follows： Tibetan Plateau, Northwest China, North China, and South China. Strong seasonal variation of the PBLH exhibited high values in summer and low values in winter, which was also consistent with the turbulent vertical exchange. Not surprisingly, the highest RH in South China and the lowest RH in desert areas of Northwest China （less than 30%）. Seasonally, South China exhibited little variation, whereas Northwest China exhibited its highest humidity in winter and lowest humidity in spring, the maximum values in the other regions were obtained from July to September.展开更多
Learning comprehensive spatiotemporal features is crucial for human action recognition. Existing methods tend to model the spatiotemporal feature blocks in an integrate-separate-integrate form, such as appearance-and-...Learning comprehensive spatiotemporal features is crucial for human action recognition. Existing methods tend to model the spatiotemporal feature blocks in an integrate-separate-integrate form, such as appearance-and-relation network(ARTNet) and spatiotemporal and motion network(STM). However, with blocks stacking up, the rear part of the network has poor interpretability. To avoid this problem, we propose a novel architecture called spatial temporal relation network(STRNet), which can learn explicit information of appearance, motion and especially the temporal relation information. Specifically, our STRNet is constructed by three branches,which separates the features into 1) appearance pathway, to obtain spatial semantics, 2) motion pathway, to reinforce the spatiotemporal feature representation, and 3) relation pathway, to focus on capturing temporal relation details of successive frames and to explore long-term representation dependency. In addition, our STRNet does not just simply merge the multi-branch information, but we apply a flexible and effective strategy to fuse the complementary information from multiple pathways. We evaluate our network on four major action recognition benchmarks: Kinetics-400, UCF-101, HMDB-51, and Something-Something v1, demonstrating that the performance of our STRNet achieves the state-of-the-art result on the UCF-101 and HMDB-51 datasets, as well as a comparable accuracy with the state-of-the-art method on Something-Something v1 and Kinetics-400.展开更多
基金supported by the National Key R&D Program of China (2016YFC0201507)the National Natural Science Foundation of China (Grant Nos. 41471367, 91543128 and 41571417)
文摘Few studies have specifically focused on the validation and spatiotemporal distribution of planetary boundary layer height （PBLH） and relative humidity （RH） data in China. In this analysis, continuous PBLH and surface-level RH data simulated from GEOS-5 between 2004 and 2012, were validated against ground-based observations. Overall, the simulated RH was consistent with the statistical data from meteorological stations, with a correlation coefficient of 0.78 and a slope of 0.9. However, the simulated PBLH was underestimated compared to LIDAR data by a factor of approximately two, which was primarily because of poor simulation in late summer and early autumn. We further examined the spatiotemporal distribution characteristics of two factors in four regions--North China, South China, Northwest China, and the Tibetan Plateau. The results showed that the annual PBLH trends in all regions were fairly moderate but sensitive to solar radiation and precipitation, which explains why the PBLH values were ranked in order from largest to smallest as follows： Tibetan Plateau, Northwest China, North China, and South China. Strong seasonal variation of the PBLH exhibited high values in summer and low values in winter, which was also consistent with the turbulent vertical exchange. Not surprisingly, the highest RH in South China and the lowest RH in desert areas of Northwest China （less than 30%）. Seasonally, South China exhibited little variation, whereas Northwest China exhibited its highest humidity in winter and lowest humidity in spring, the maximum values in the other regions were obtained from July to September.
基金supported by National Natural Science Foundation of China(Nos.U1836218,62020106012,61672265 and 61902153)the 111 Project of Ministry of Education of China(No.B12018)+1 种基金the EPSRC Programme FACER2VM(No.EP/N007743/1)the EPSRC/MURI/Dstl Project under(No.EP/R013616/1.)。
文摘Learning comprehensive spatiotemporal features is crucial for human action recognition. Existing methods tend to model the spatiotemporal feature blocks in an integrate-separate-integrate form, such as appearance-and-relation network(ARTNet) and spatiotemporal and motion network(STM). However, with blocks stacking up, the rear part of the network has poor interpretability. To avoid this problem, we propose a novel architecture called spatial temporal relation network(STRNet), which can learn explicit information of appearance, motion and especially the temporal relation information. Specifically, our STRNet is constructed by three branches,which separates the features into 1) appearance pathway, to obtain spatial semantics, 2) motion pathway, to reinforce the spatiotemporal feature representation, and 3) relation pathway, to focus on capturing temporal relation details of successive frames and to explore long-term representation dependency. In addition, our STRNet does not just simply merge the multi-branch information, but we apply a flexible and effective strategy to fuse the complementary information from multiple pathways. We evaluate our network on four major action recognition benchmarks: Kinetics-400, UCF-101, HMDB-51, and Something-Something v1, demonstrating that the performance of our STRNet achieves the state-of-the-art result on the UCF-101 and HMDB-51 datasets, as well as a comparable accuracy with the state-of-the-art method on Something-Something v1 and Kinetics-400.
