为了满足田间作物长势快速检测与指导变量管理的需求,以玉米为例设计了基于多光谱成像的田间作物植株叶绿素检测系统,包括可见光(RGB)和近红外(Near-infrared,NIR)图像采集模块、主控处理器模块、模型加速模块、显示及电源模块,用于实...为了满足田间作物长势快速检测与指导变量管理的需求,以玉米为例设计了基于多光谱成像的田间作物植株叶绿素检测系统,包括可见光(RGB)和近红外(Near-infrared,NIR)图像采集模块、主控处理器模块、模型加速模块、显示及电源模块,用于实现玉米植株智能识别与叶绿素指标一体化检测。首先,采集玉米苗期和拔节期冠层图像数据集,比较了植株冠层实例分割与株心目标检测两种深度学习模型,构建了基于MobileDet+SSDLite(Single shot multibox detector lite)轻量化网络的玉米植株定位检测模型,实现玉米植株识别。其次,提取被识别的植株株心RGB-NIR图像,开展RGB和NIR图像匹配与分割,提取R、G、B和NIR灰度值计算植被指数,使用SPXY算法(Sample set portioning based on joint X-Y distances)和连续投影算法(Successive projections algorithm,SPA)分别对数据集进行样本划分及特征变量筛选,选择高斯过程回归(Gaussian process regression,GPR)算法建立叶绿素指标检测模型。结果显示,玉米株心目标检测模型在遮挡重叠的复杂环境下识别率达到88.7%,在不交叉重叠时识别精度达到90%以上;叶绿素含量指标检测模型建模集的模型决定系数R^(2)为0.62,测试集模型决定系数R^(2)为0.61。对开发系统进行田间测试,结果显示,系统检测速率可达14.6 f/s,平均精度为92.9%。研究结果能够有效解决大田环境下玉米营养状态的检测问题,满足大田环境实时检测需求,为作物生产智慧感知提供解决思路和技术支持。展开更多
Study on Chlorophyll a （Chl. a） distribution and bloom patterns is essential in the upwelling areas that constitute the main fishery grounds of Vietnam. Based on high resolution satellite imagery and monthly in-situ...Study on Chlorophyll a （Chl. a） distribution and bloom patterns is essential in the upwelling areas that constitute the main fishery grounds of Vietnam. Based on high resolution satellite imagery and monthly in-situ data in period from 2007 to 2008, the spatial and temporal variations ofChl, a in the upwelling region including algal blooms have been detected. Anomalies of higher Chl. a concentration and bloom patterns occurred at different periods, locations, and bloom shapes in coastal waters of Southern Center of Vietnam. The appearances of these bloom are related to four difference reasons, such as： （i） Their locations coincided to main center of upwelling phenomena that usually occur during southwest monsoon （SWM）; （ii） Effect of northeast monsoon （NEM） that brings and attaches the nutrient into the coast and created ＂floating＂ algae bloom patches in coastal zones; （iii） The algae blooms can be potentially associated with Harmful Algae Bloom （HAB） during the SWM or local eddies during NEM and （iv） Their appearances are also as a result of distinctive nutrient inflow fi＇om Mekong delta and pumping of nutrients by internal waves in coastal shallow waters.展开更多
文摘为了满足田间作物长势快速检测与指导变量管理的需求,以玉米为例设计了基于多光谱成像的田间作物植株叶绿素检测系统,包括可见光(RGB)和近红外(Near-infrared,NIR)图像采集模块、主控处理器模块、模型加速模块、显示及电源模块,用于实现玉米植株智能识别与叶绿素指标一体化检测。首先,采集玉米苗期和拔节期冠层图像数据集,比较了植株冠层实例分割与株心目标检测两种深度学习模型,构建了基于MobileDet+SSDLite(Single shot multibox detector lite)轻量化网络的玉米植株定位检测模型,实现玉米植株识别。其次,提取被识别的植株株心RGB-NIR图像,开展RGB和NIR图像匹配与分割,提取R、G、B和NIR灰度值计算植被指数,使用SPXY算法(Sample set portioning based on joint X-Y distances)和连续投影算法(Successive projections algorithm,SPA)分别对数据集进行样本划分及特征变量筛选,选择高斯过程回归(Gaussian process regression,GPR)算法建立叶绿素指标检测模型。结果显示,玉米株心目标检测模型在遮挡重叠的复杂环境下识别率达到88.7%,在不交叉重叠时识别精度达到90%以上;叶绿素含量指标检测模型建模集的模型决定系数R^(2)为0.62,测试集模型决定系数R^(2)为0.61。对开发系统进行田间测试,结果显示,系统检测速率可达14.6 f/s,平均精度为92.9%。研究结果能够有效解决大田环境下玉米营养状态的检测问题,满足大田环境实时检测需求,为作物生产智慧感知提供解决思路和技术支持。
文摘Study on Chlorophyll a （Chl. a） distribution and bloom patterns is essential in the upwelling areas that constitute the main fishery grounds of Vietnam. Based on high resolution satellite imagery and monthly in-situ data in period from 2007 to 2008, the spatial and temporal variations ofChl, a in the upwelling region including algal blooms have been detected. Anomalies of higher Chl. a concentration and bloom patterns occurred at different periods, locations, and bloom shapes in coastal waters of Southern Center of Vietnam. The appearances of these bloom are related to four difference reasons, such as： （i） Their locations coincided to main center of upwelling phenomena that usually occur during southwest monsoon （SWM）; （ii） Effect of northeast monsoon （NEM） that brings and attaches the nutrient into the coast and created ＂floating＂ algae bloom patches in coastal zones; （iii） The algae blooms can be potentially associated with Harmful Algae Bloom （HAB） during the SWM or local eddies during NEM and （iv） Their appearances are also as a result of distinctive nutrient inflow fi＇om Mekong delta and pumping of nutrients by internal waves in coastal shallow waters.
