基于IMU的细粒度奶牛行为判别

针对奶牛行为判别自动化水平不足、准确率低的问题,采用惯性测量单元(IMU)和卷积神经网络(CNN),对细粒度奶牛行为判别进行研究.结果表明:1)在KNN、SVM、BPNN、CNN和LSTM 5个模型中,CNN模型在奶牛行为分类测试集上的准确率最高.2)含有三轴加速度计、陀螺仪和磁力计的IMU更加适用于奶牛行为分类,其分...     

拖拉机自动驾驶系统的研究

主要论述拖拉机自动驾驶在农业中应用的必要性以及国内外拖拉机自动驾驶的发展,依据学校211重点实验室建设项目针对铁牛654拖拉机搭建拖拉机自动驾驶平台,论述了平台的硬件组成及工作原理和基于GPS/IMU定位与导航的控制策略。     

Stereo vision three-dimensional terrain maps for precision agriculture

The combined interest in precision agriculture, information technology, and autonomous navigation has led to a growing interest in the generation of 3D maps of mobile equipment surroundings. This article proposes a method to create 3D terrain maps by combining the information captured with a stereo camera, a localization sensor, and an inertial measurement unit, all installed on a mobile equipment platform. The perception engine comprises a compact stereo camera that captures field scenes and generates 3D point clouds, which are transformed to geodetic coordinates and assembled in a global field map. The results showed that stereo perception can provide the level of detail and accuracy needed in the construction of 3D field maps for precision agriculture and field robotics applications.     

果园自主导航兼自动对靶喷雾机器人

为同时实现果园智能植保机自主导航及自动对靶喷雾,研制了一种果园自主导航兼自动对靶喷雾机器人。首先采用单个3D LiDAR（Light Detection and Ranging）采集果树信息确定兴趣区（Region of Interest,ROI）,对ROI内点云进行2D化处理得到果树质心坐标,通过随机一致性（Random Sample Consensus,RANSAC）算法得到果树行线,并确定果树行中间线（导航线）,进而控制机器人沿导航线行驶。通过编码器及惯性测量单元（Inertial Measurement Unit,IMU）确定机体速度及位置,IMU矫正采集到的果树分区冠层信息,最后通过程序判断分区冠层的有无控制喷头是否喷雾。结果表明,机器人自主导航时最大横向定位偏差为21.8 cm,最大航向偏角为4.02°,相比于传统连续喷雾机施药液量、空中漂移量及地面流失量分别减少20.06%、38.68%及51.40%。本研究通过单个3D LiDAR、编码器及IMU在保证喷雾效果的前提下,实现了喷雾机器人自主导航及自动对靶喷雾,降低了农药使用量及飘失量。     

激光扫描仪结合GPS/IMU在地面三维形状测量中的应用

阐述了开发使用比全站仪更为灵活的车载GPS/IMU(全球定位系统/惯性导航系统)结合地面型三维激光扫描仪,在地面三维形状测量的应用,并就系统构成,数据解析、检证、量测精度等进行了探讨。     

Vibration analysis during grass harvesting according to ISO vibration standards

This research evaluated the working efficiency and comfort of operation by measuring vibration acceleration of tractors during grass harvesting. A real-time kinematic global positioning system and an inertial measurement unit installed in a tractor normally used by farmers during grass harvesting were used to acquire tractor vibration acceleration data. Analysis of the position and vibration acceleration data of tractors by a Fourier transformation yielded a power spectrum of vibration acceleration at each frequency (1-10 Hz) and position. The root mean square of vibration acceleration at each frequency (1-10 Hz) was calculated with the center frequency of the 1/3 octave bands (1.0, 1.25, 1.6, 2.0, 2.5, 3.15, 4, 5, 6.3, 8, and 10 Hz) based on ISO standards. To evaluate the working efficiency in the grassland, geographical information system maps were generated using the power spectrum of vibration acceleration and the limit on working time for each frequency that negatively affected the tractor driver. The vibration acceleration in the longitudinal (a_x) and lateral (a_y) directions at the center frequency of the 1/3 octave band below 2.0 Hz exceeded the fatigue-decreased proficiency and reduced the comfort boundaries stipulated in ISO 2631 (1974). In the area where working characteristics are severe, the vibration acceleration in the vertical direction (a_z) is high. The vibration acceleration in the a_z direction at the center frequency of the 1/3 octave band (5-10 Hz) clearly indicates discomfort during grass harvesting and a decrease in the work efficiency beyond 1 h. The total vibration acceleration (a_ν) at the center frequency of 5.0 Hz of the 1/3 octave band served to evaluate comfort in the whole field during grass harvesting—the a_ν value is higher than that at other frequencies. The area with severe working characteristics showed a higher a_ν value at the center frequency of 8.0 Hz than that at other frequencies.     

果园机器人LiDAR/IMU紧耦合实时定位与建图方法

针对果园环境中GNSS定位信号易丢失和传统SLAM算法鲁棒性较差的问题，本文提出一种基于LiDAR/IMU紧耦合框架的全局无偏状态估计果园机器人定位与建图方法。LiDAR/IMU紧耦合框架基于因子图进行多源约束的IMU里程计构建，实时输出高频位姿信息，IMU里程计因子和预积分因子优化LiDAR里程计并提供位姿先验约束IMU零偏。引入局部点云地图参与特征点云粗匹配和非特征点云递进式匹配进一步稠密化源点云，改善LiDAR里程计的性能。融合GPS信号与LiDAR/IMU紧耦合框架的地图构建，能够得到准确且高频连续的位姿信息，提高点云地图的复用率。在果园和苗木等场景验证了该算法的性能，实验结果表明，与LIO-SAM等算法相比，定位精度维持在0.05 m左右，均方根误差为0.016 2 m。本文算法使机器人具有更高的精度、实时性和鲁棒性，有效降低了系统累积误差，保证了所构建地图的全局一致性。     

基于BDS和IMU的挖掘机铲斗位姿测量方法与试验

为提高农田建设中挖掘机施工作业精度和智能化程度，本文提出了一种基于BDS（BeiDou Navigation Satellite System，北斗卫星导航系统）和IMU（Inertial Measurement Unit，惯性测量单元）的挖掘机铲斗位姿测量方法。首先，采用IMU测量挖掘机各执行机构的姿态角信息，解算获得挖掘机车体坐标系下铲斗末端的三维坐标，利用双天线BDS和IMU检测车体的位置和姿态建立了挖掘机铲斗末端三维坐标的实时解算模型，并设计了融合双天线BDS和IMU输出高频率高精度位姿的卡尔曼滤波算法。模拟挖掘机实际施工场景进行了静态和动态试验，采用全站仪验证铲斗末端三维坐标解算值。试验结果表明，该方法能准确实时测量挖掘机铲斗末端三维坐标，挖掘机铲斗末端三维坐标解算值与全站仪实测值的运动轨迹变化一致，同一时刻空间两坐标点距离均方根偏差小于30mm，三个轴向坐标动态测量均方根偏差均在20mm内，绝对偏差≤30mm的数据占比不低于95.35%，挖掘机铲斗位姿的准确测量为挖掘机精准施工智能引导提供了基础。