摘要
本文研究了无人机倾斜摄影在农村不动产确权测量过程中的技术革新策略,分析了农村不动产测量的任务,同时分析了无人机在农村区域的空域环境,发现应通过机器学习方法解决局部湍流带来的无人机倾斜摄影图像的平行四边形畸变和梯形畸变。使用卷积多列神经网络基于无人机机载设备(陀螺仪、加速度计、GPS等)提供的融合数据,对OpenSVG位图引擎提供图像修正参数,将修正的数据通过Smart 3D工具包软件进行后续分析。在不改变传统方案其他工作流要素的情况下,通过融入上述图像数据预处理过程,使测量结果精细化程度增加,村民代表对测量结果的主观感性评价也得到提升,即该技术革新不但具有工程价值,还具备一定的社会价值。
This paper studies the technical innovation strategy of UAV tilt photography in the process of rural real estate ownership measurement.By analyzing the task of rural real estate measurement and the airspace environment of UAV in rural area,it is found that the parallelogram distortion and trapezoid distortion of UAV oblique photography image caused by local turbulence should be solved by machine learning method.The convolution multi column neural network is used to provide image correction parameters for OpenSVG bitmap engine based on the fusion data provided by UAV airborne equipment(gyroscope,accelerometer,GPS,etc.),and the corrected data is analyzed through Smart 3D toolkit software.Without changing other workflow elements of the traditional scheme,by integrating the above image data process,the degree of refinement of the measurement results is increased,and the subjective perceptual evaluation of the measurement results by villagers′representatives is also improved.That is,the technological innovation not only has engineering value,but also has certain social value.
作者
周新星
袁伟俭
陈曼曼
ZHOU Xinxing;YUAN Weijian;CHEN Manman(Zhejiang Provincial Land Survey and Planning Co.,Ltd.,Wenzhou 325000,China;Zhejiang Academy of Surveying&Mapping,Wenzhou 325000,China;Wenzhou Ouhai Institute of Surveying,Mapping and Geographic Information,Wenzhou 325000,China)
出处
《测绘与空间地理信息》
2021年第10期204-207,共4页
Geomatics & Spatial Information Technology
关键词
无人机
倾斜摄影
农村不动产调查
确权测量
机器学习
UAV
tilt photography
rural real estate survey
right confirmation measurement
machine learning
作者简介
周新星(1992-),男,浙江宁波人,助理工程师,本科学历,主要从事工程测量方面的工作。
