| 摘要: |
| 针对使用智能手机野外测量叶面积时很难保证主光轴与叶平面垂直导致测量精度下降的问题，采用针孔相机模型和数字图像处理技术相结合的方法，对叶片平面和手机平面存在夹角时的叶面积测量建立误差校正模型并进行研究。1）以10片不同大小的椭圆形模拟叶片对开发的android App进行仿真测试，结果显示，叶面积测量值与理论值之间的决策系数高于97%，并且随着叶片与手机平面夹角的增大而减小，校正后的相对误差为-6%~8%，未校正的高达41.4%。2）以向日葵、番茄、辣椒和茄子等植物叶片对误差校正公式进行验证，结果显示，当夹角<15°，拍照高度为30~65 cm时，叶面积测量的相对误差为-6%~6%，未校正的为-27.5%~25%。 |
| 关键词: 叶面积 针孔相机模型 android App 误差校正 |
| DOI：10.11841/j.issn.1007-4333.2017.06.15 |
| 投稿时间：2016-05-31 |
| 基金项目:中央高校基本科研业务费专项资金（2015ZCQ-GX-03）；国家自然科学基金面上项目（31371537） |
|
| Leaf area measurement correction based on smartphone orientation sensor |
|
GAO Dashuai1, LIN Jianhui1, YU Jiangbin1, SUN Yurui2, SCHULZE Lammers Peter3
|
| (1.School of Technology, Beijing Forestry University, Beijing 100083, China;2.School of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China;3.Institute of Agricultural Engineering, University of Bonn, Bonn 53115, Germany) |
| Abstract: |
| Aiming at the problem that it is difficult to keep phone camera's optical axis vertical with leaf plane when measuring leaf area under field condition.The pinhole camera model and digital image processing techniques were used to correct the leaf area measurement in this study to build an error correction model for the quantitative influence of camera tilt angle factor on leaf area measurement.The results showed that:1) Ten virtual oval green leafs with different sizes were used to test the developed android APP.Simulation results showed that the decision coefficients between measured area and theoretical area were higher than 97% and decrease with the increase of the camera tilt angle,the relative error of the corrected leaf area was range from -6% to 8%,while the uncorrected error was up to 41%.2) Leaves of sunflower,tomato,pepper,and eggplants were selected to verify the proposed method.Experiment results showed that when the camera tilt angle was less than 15 degrees and the distance between leaf and phone was about 30 to 65 cm,the measured relative error was ranged from -6% to 6%,while the uncorrected error was ranged from -27.5% to 25%. |
| Key words: leaf area pinhole camera model android App error correction |
