PHANTOM 4 RTK+大疆像控处理技术在燕麦长势模拟中的应用

PHANTOM 4 RTK+ DJI image processing technology in oat growth simulation application

关键词:  燕麦  无人机  高度  快速  高精度

Key words:  oats  UAV  height  fast  high precision

摘要:

【目的】 利用小型消费级无人机航拍获取地物影像，通过地物阴影、高度差、色差快速提取地物，进而获取地物结构信息。【方法】 文章选取云南省曲靖市会泽县的大桥乡为研究区域，针对冬闲田闲置土地资源、种植结构相对单一的区域展开试验，利用高分辨率无人机遥感影像对燕麦进行识别，同时结合超声波传感器数据估算地物高度，并与实际高度和无人机生成的传统测高方法得到的高度进行相关性分析，获取高精度、可靠性强的数据。【结果】 基于可见光燕麦的总体分类精度为91.46%，Kappa系数为0.857，在增加DSM数据后的分类总体精度为98.91%，Kappa系数为0.982。研究表明由无人机获取的代表燕麦冠层高度信息的DSM数据能够显著提升燕麦的识别效果。相对于传统无人机测高方法生成数字表面模型提取地物高度的方法，依赖于光谱和高程信息识别地物信息的方法在计算地物高度时，精度更高，识别结果更可靠。【结论】 该文提出的小型消费级无人机利用地物阴影计算燕麦高度的方法，改进了相机镜头光心地位和RTK天线中心点地位补偿作用，打通了RTK模块、飞控模块及相机云台模块之间的通讯，能够应用于实际准确获取影像地位信息，为无人机遥感快速、准确地获取地物高度信息提供了一种新的思路。

Abstract:

[Purpose] Using small consumer UAV aerial photography to acquire ground object image，quickly extract ground object through ground object shadow，height difference，and color difference，and then obtain ground object structure information.[Method] In this study，Huize County in the Wumeng Mountain area was selected as the research area，and the experiment was carried out in the area where the idle land resources of winter fallow field and the planting structure was relatively single. The high-resolution UAV remote sensing image was used to identify oat，and the height of the ground object was estimated combined with ultrasonic sensor data. The correlation analysis with the actual height and the height obtained by the traditional altimetry method generated by UAV is carried out to obtain high-precision and reliable data.[Result] The experimental results showed that the overall classification accuracy of oat-based on visible light was 91.46%，and the Kappa coefficient was 0.857. After adding DSM data，the overall classification accuracy was 98.91%，and the Kappa coefficient was 0.982. The study showed that the DSM data of oat canopy height obtained by UAV could significantly improve the recognition effect of oat. Compared with the traditional UAV altimetry method，which generates a digital surface model to extract ground object height，the ground object recognition method relying on spectral and elevation information has higher accuracy and more reliable recognition results in calculating ground object height.[Conclusion] In this study，the method of calculating oat height by ground object shadow for small consumer UAV improves the compensation effect of the optical center position of camera lens and RTK antenna center position，and the communication between RTK module，flight control module，and camera pall module can be applied to obtain image position information accurately in practice. It provides a new idea for UAV remote sensing to acquire ground object height information quickly and accurately.