Research On UAV Data Acquisition And Processing For Industrial Parks

In recent years,the topic of smart industrial parks has become more and more heated.The three-dimensional modeling of it can achieve functions such as multi-data linkage,real-time perception of the park situation,and three-dimensional comprehensive display.It can provide powerful technical support for various departments when planning,and save Eliminate unnecessary consumption of human and financial resources.Traditional 3D modeling mostly uses ground-based laser instruments to collect data on targets.When the scale of the collected objects increases,problems such as higher modeling costs and difficulty in internal data processing often occur.However,the 3D modeling technology of tilt photogrammetry based on drones can effectively solve these problems,quickly and accurately obtain high-resolution image data,and then obtain high-precision 3D models through internal processing,which has very broad research prospects.In this paper,UAV is used as a remote sensing data acquisition platform,combined with tilt photogrammetry technology to carry out 3D modeling of Quanzhou Robot Industrial Park.The thesis conducts research from three aspects:UAV field data collection,UAV image defogging processing,and internal data processing to generate three-dimensional models.The specific contents are as follows:(1)Through on-site exploration and project requirements,the drone and camera models were determined.Based on this,a UAV data acquisition system was built with the Haikang Eagle series MX6150 B UAV as the flying platform and the G5V2020 A five-lens camera,which realized the data collection function of Quanzhou Robot Industrial Park.(2)A detailed study of the UAV field data collection process,that is,the UAV image acquisition process.Among them,based on the characteristics of Quanzhou Robot Industrial Park,the traditional image control layout scheme is improved,and a set of optimal image control layout schemes suitable for the application of this article in Quanzhou Robot Industrial Park are designed.(3)The Retinex defogging algorithm and the dark channel a priori defogging algorithm are studied in detail.The dark channel a priori defogging algorithm is selected as the main research algorithm in this paper.The traditional dark channel a priori defogging algorithm is optimized by improving the calculation of the atmospheric light A value,and the experimental verification is completed.Finally,this article uses Quanzhou Robot Industrial Park as an experimental area to conduct research.UAV data acquisition system is used to obtain image data of the experimental area,and Pix4 Dmapper software is used to construct a three-dimensional model of the park.Improved the accuracy of the three-dimensional model of the industrial park by improving the image control layout and defogging the UAV image.Through field measurement in the Quanzhou Robot Industrial Park,the control point coordinates,building height and other parameter information were obtained,combined with the actual measurement values of the model,the accuracy analysis was completed,and it was proved that the 3D model generated in this paper conforms to the?3D Geographic Information Model Data Product Specification?,can be applied to actual production.