To investigate the nitrifying activities of different soil types, soil samples collected from 8-, 50- and 90-year old tea orchards, the adjacent wasteland, and 90-year old forest were measured for their nitrification ...To investigate the nitrifying activities of different soil types, soil samples collected from 8-, 50- and 90-year old tea orchards, the adjacent wasteland, and 90-year old forest were measured for their nitrification potentials using the conventional soil incubation and the liquid incubation method. Among different soil types, the nitrification potential of soil in tea orchards was higher than that of wasteland and forest soils. The slurry shaken liquid incubation method was confirmed to be more accurate and have reliable results than the soil incubation. Interestingly, experimental result revealed that the generally applied pH value of 7.2 for the liquid media was not the optimal pH for these acid soils with a strong buffer capacity. This suggested that tea orchard soils may have nitrifiers requiring pHneutral condition for the best activity. Our data also showed that treatment with the commonly used nitrogen fertilizer urea significantly improved nitrification potential of the soils; such enhancement effect was stronger on all of three tea orchard soils than on wasteland and forest soils, and also stronger on the younger (8- and 50-year old) tea orchard soils than on the older one (90-year old).展开更多
针对果、茶园规模不断扩张并逐渐向智能农业机械化发展的趋势以及常用道路语义分割数据集缺少果、茶园道路场景等问题,将语义分割技术应用到部分果、茶园道路中,以实现对果、茶园道路的像素级分割。以道路、人和车为分类对象,建立果、...针对果、茶园规模不断扩张并逐渐向智能农业机械化发展的趋势以及常用道路语义分割数据集缺少果、茶园道路场景等问题,将语义分割技术应用到部分果、茶园道路中,以实现对果、茶园道路的像素级分割。以道路、人和车为分类对象,建立果、茶园道路场景图像数据集(包括6032张图像),将数据集按照9∶1比例随机划分为训练集(5429张图像)和测试集(603张图像)。以PSPNet(pyramid scene parsing network,金字塔场景解析网络)分割模型为基础进行优化,构建MS-PSPNet语义分割模型;训练结果显示,MS-PSPNet模型的MIoU(mean intersection over union,平均交并比)为83.41%,FPS(frames per second,每秒传输帧数)为22.31。将MS-PSPNet模型应用在果、茶园不同道路条件和光照强度下进行现场试验,并进行准确度评估,结果显示,MS-PSPNet模型类别MPA(mean pixel accuracy),像素准确率均超过92%,MIoU在除非硬化道路条件情况均超过91%,表明MS-PSPNet模型在果、茶园道路识别中具有较好的有效性和适用性。展开更多
基金supported by the National Natural Science Foundation of China(No.30671207,30871600)Zhejiang Provincial National Natural Science Foundation of China(No.Y5080067)the Doctoral Scientific Research Foundation of Luoyang Institute of Science and Technology(No.2008BZ04)
文摘To investigate the nitrifying activities of different soil types, soil samples collected from 8-, 50- and 90-year old tea orchards, the adjacent wasteland, and 90-year old forest were measured for their nitrification potentials using the conventional soil incubation and the liquid incubation method. Among different soil types, the nitrification potential of soil in tea orchards was higher than that of wasteland and forest soils. The slurry shaken liquid incubation method was confirmed to be more accurate and have reliable results than the soil incubation. Interestingly, experimental result revealed that the generally applied pH value of 7.2 for the liquid media was not the optimal pH for these acid soils with a strong buffer capacity. This suggested that tea orchard soils may have nitrifiers requiring pHneutral condition for the best activity. Our data also showed that treatment with the commonly used nitrogen fertilizer urea significantly improved nitrification potential of the soils; such enhancement effect was stronger on all of three tea orchard soils than on wasteland and forest soils, and also stronger on the younger (8- and 50-year old) tea orchard soils than on the older one (90-year old).
文摘针对果、茶园规模不断扩张并逐渐向智能农业机械化发展的趋势以及常用道路语义分割数据集缺少果、茶园道路场景等问题,将语义分割技术应用到部分果、茶园道路中,以实现对果、茶园道路的像素级分割。以道路、人和车为分类对象,建立果、茶园道路场景图像数据集(包括6032张图像),将数据集按照9∶1比例随机划分为训练集(5429张图像)和测试集(603张图像)。以PSPNet(pyramid scene parsing network,金字塔场景解析网络)分割模型为基础进行优化,构建MS-PSPNet语义分割模型;训练结果显示,MS-PSPNet模型的MIoU(mean intersection over union,平均交并比)为83.41%,FPS(frames per second,每秒传输帧数)为22.31。将MS-PSPNet模型应用在果、茶园不同道路条件和光照强度下进行现场试验,并进行准确度评估,结果显示,MS-PSPNet模型类别MPA(mean pixel accuracy),像素准确率均超过92%,MIoU在除非硬化道路条件情况均超过91%,表明MS-PSPNet模型在果、茶园道路识别中具有较好的有效性和适用性。