红外相机技术在河北雾灵山国家级自然保护区中的应用

为了充实河北雾灵山自然保护区的野生动物资源本底资料,提高监测能力和监测水平,雾灵山保护区自2015年开始将红外相机技术应用到野生动植物的资源调查和管护中。目前应用红外相机技术发现兽类5目10科14种,鸟类4目6科10种,掌握了区内主要物种猪獾(Arctonyx collaris)、狍(Capreolus)和勺鸡(Pucrasia macrolopha)的日活动节律,补充了保护区的本底调查资料,为保护区进一步开展金钱豹(Panthera pardus)、斑羚(Naemorhedus goral)和鸟类多样性等科学研究奠定了基础,为保护区制定关于野生动物资源的相关管护政策提供了科学依据。     

Determining spatial feed distribution in sea cage aquaculture using an aerial camera platform

Within this paper we describe an unmanned aerial vehicle (UAV) based method to estimate the spatial feed pellet distribution in salmon fish-cages and exploit it within a case study to determine the radial pellet distribution for different rotor spreaders and blower configurations. Compared to previously used methods, capturing pellets thrown from a rotary feed spreader in rows of Styrofoam boxes, the UAV based method is simpler and faster to setup and allows to cover a larger portion of the sea cage surface area. We compare results obtained with the Styrofoam box method with results we obtained by an automatic analysis of aerial videos taken by the UAV during feeding experiments. The employed method helps to gain insight into the spatial feed pellet distribution in full-scale salmon fish-cages where the feeding pipe and actual fixation of the spreader may influence the dynamic behaviour of the feeding system. This can also be seen as a step towards a continuous measuring of the feed pellet distribution, particularly in view of possible future feeding systems that may allow to adapt the feed pellet distribution to the actual fish distribution.     

基于改进SIFT-ICP算法的Kinect植株点云配准方法

针对传统配准方法准确度低、速度慢的问题,提出了基于改进SIFT-ICP算法的彩色植株点云配准方法。首先采用Kinect获取不同视角下植株彩色图像和深度图像合成原始植株彩色点云,通过预处理提取原始点云植株信息,对植株点云进行尺度不变特征变换(SIFT)的特征点检测,得到点云配准关键点,再对关键点进行自适应法线估计,然后求取关键点的快速点特征直方图(FPFH),通过采样一致性(SAC-IA)初始配准方法改进点云间初始位置关系,最后利用Nanoflann加速最近点迭代(ICP)算法完成精确配准。试验结果表明,改进SIFT-ICP算法可以大幅度提高点云配准的准确性和快速性,其中对应点间平均欧氏距离小于7 mm,配准时间小于30 s。     

基于Kinect动态手势识别的机械臂实时位姿控制系统

基于Kinect动态手势识别达到实时控制机械臂末端位姿的效果。位置控制信息的获取采用Kinect计算手部4个关节点在控制中的位置变动,数据噪声在控制中易引起机械臂误动作和运动振动等问题,为了避免噪声对实时控制的不利影响,采用卡尔曼滤波跟踪降噪。姿势控制信息通过采集手部点云经滤波处理后应用最小二乘拟合的方式获取掌心所在平面,运用迭代器降噪处理。系统通过对手部位置和姿势信息的整合、手势到机械臂空间坐标映射及运动学求解来实时控制机械臂末端位姿。实验结果证明,手势控制系统满足控制要求,简单、易于操作,机械臂实时响应速度快、运动准确。     

喷杆式施药机对行喷雾控制系统设计与试验

针对现有大田喷杆式施药机喷雾过程中喷头无法精准对行喷施造成农药浪费的问题,基于机器视觉技术设计了喷杆式施药机对行喷雾控制系统。该系统包括作物行中心线位置提取上位机软件和电动喷杆控制系统,利用工业相机获取作物行RGB图像,采用G-RTG-BT算法及形态学处理实现作物行分割,基于改进的垂直投影法获取作物行中心线,利用坐标系转换实现将作物行中心线位置信息转化为喷杆横向偏移量,并经RS2 3 2串口传输至ATMega1 6控制器,控制推杆电机带动喷杆在滑轨上左右移动,借助位移传感器实时监测喷杆移动距离,以实现作物行追踪和对行喷雾控制。实验室和田间试验表明:改进的作物行中心线提取算法平均耗时12.51ms,喷杆横向偏移量计算误差小于0.44cm;电动喷杆右移最大误差0.3cm,左移最大误差0.5cm;小车速度为0.26m/s时,对倾角为5°、10°、15°模拟作物行的最大对行误差分别为3.22、2.86、2.51cm;小车速度为0.2 4 m/s,最大偏移1 4.0 2 cm时,对田间玉米幼苗的对行喷雾最大误差为4.8 6 cm,为实现作物行追踪和对行喷雾控制提供了一种有效的解决方案。     

Alpha-shape算法构建枣树点云三维模型

为了实现枣树智能化修剪作业，该研究提出了基于点云配准的自然光照环境下的果树三维重构方法，并针对传统最近点迭代（Iterative Closest Point，ICP）算法对待配准点云的空间位置要求苛刻的问题，提出了改进的点云配准算法。首先，使用彩色深度（RGB-D）相机采集不同角度下的枣树彩色和深度图像，并通过信息融合实现相应角度下的点云获取。其次，对点云进行背景去除和滤波处理，基于直方图设定分割阈值，提取单株枣树点云，并将放置在树根附近的标靶球作为标记，使用人工标记法进行两站点云初配准。最后，在初配准基础上计算点云的曲面法向量和曲率，由曲率相近的点构成配对点对，使用k维树最近点迭代（k dimensional-tree-Iterative Closest Point，kd-tree-ICP）算法完成精配准，对点云使用Alpha-shape算法面片化，实现表面重构。利用上述方法对多棵枣树进行全局配准并完整重构果树模型。试验结果表明，通过引入初配准，有效提高了点云配准的准确性和稳定性，配准误差均控制在1.0 cm以内，平均配准误差为0.78 cm；重构模型真实感较强，在外观上更加接近真实树，枝干相对误差控制在7%以内。该研究重构模型精度较高，可为枣树智能修剪提供可视化研究基础和技术支持。     

基于深度相机的玉米株型参数提取方法研究

提出了一种基于骨架提取的改进算法,可实现在大田环境下,使用PMD深度相机快速、无损测量玉米株型参数。首先利用深度图像RGB伪彩色和深度距离信息,提取深度图像的骨架,排除复杂背景干扰,得到单株玉米的二值骨架图像;然后利用基于角点检测的改进归类算法提取骨架图像特征点;最后建立骨架图像中特征点与深度图像的对应关系,利用空间几何数学方法,结合特征点计算出玉米的3种株型参数,即株高、茎粗、叶倾角。农田实验对比分析表明,所提方法的株高测量结果与人工测量结果的相关系数 r 为0.986,最大相对误差小于2 cm,农田作物育种抗逆性分析还表明玉米株型参数与抗倒伏性具有显著相关性。     

Remote sensing of biotic and abiotic stress for irrigation management of cotton

The applicability of commercially available remote sensing instrumentation was evaluated for site-specific management of abiotic and biotic stress on cotton (Gossypium hirsutum L.) grown under a center pivot low energy precision application (LEPA) irrigation system. This study was conducted in a field where three irrigation regimes (100%, 75%, and 50% ET_c) were imposed on areas of Phymatotrichum (root rot) with the specific objectives to (1) examine commercial remote sensing instrumentation for locating areas showing biotic and abiotic stress symptomology in a cotton field, (2) compare data obtained from commercial aerial infrared photography to that collected by infrared transducers (IRTs) mounted on a center pivot, (3) evaluate canopy temperature changes between irrigation regimes and their relationship to lint yield with IRTs and/or IR photography, and (4) explore the use of deficit irrigation and the use of crop coefficients for irrigation scheduling. Pivot-mounted IRTs and an IR camera were able to differentiate water stress among irrigation regimes. The IR camera distinguished between biotic (root rot) and abiotic (drought) stress with the assistance of groundtruthing. The 50% ET_c regime had significantly higher canopy temperatures than the other two regimes, which was reflected in significantly lower lint yields when compared to the 75% and 100% ET_c regimes. Deficit irrigation down to 75% ET_c had no impact on lint yield, indicating that water savings were possible without reducing yield.     

Continuous Monitoring of Visible and Near-Infrared Band Reflectance from a Rice Paddy for Determining Nitrogen Uptake Using Digital Cameras

Abstract

A two-band digital imaging system —one band for the visible red band (RED, 630?670 nm) and the other for the near infrared band (NIR, 820?900 nm)— was devised and positioned at a height of 12 m above a rice field of 300 m² in area during the 2007 growing season. The imaging system automatically logged bird’seye view images at 10-min intervals from 0800?1600 every day. Radiometric corrections for the pairs of two-band images were done using solar irradiance sensors and preceding calibrations to calculate daily band-reflectance and the normalized difference vegetation index (NDVI) values for 9 plots of rice plants, with 3 levels of planting density and basal fertilization. The daily- averaged reflectance values in the RED and the NIR bands showed different but smooth seasonal changing patterns according to the growth of plants. At the maximum tiller number and the panicle formation stages, the RED and NIR reflectance values had correlation coefficients (r) of 0.79 and 0.81 with above-ground nitrogen absorption per unit land area (NA, g m-²), respectively, whereas the NDVI using the two band reflectance values showed r-value of -0.13. An empirically derived equation for the NA using two band reflectance values showed r-value of 0.96 and a root mean square of error (RMSE) 0.5 g m^–2 (10% of the mean observed NA) in the estimation for the original (not validated) data set acquired at the maximum tiller number and the panicle formation stages. The results indicated that reflectance observation in the RED and NIR bands using the digital imaging system was potentially effective for assessing rice growth.