Research On Key Technologies Of Island And Shallow Sea Topographic Survey Based On Light And Small Photon Counting LiDAR

Airborne Li DAR,as an active optical remote sensing technology that can efficiently,quickly and accurately obtain 3D geographic information of a target,has become one of the popular applications for topographic surveying of islands above water and shallow seas.Most of the traditional lidars use a linear detection system,and the detectors have high requirements for the received laser echo signal strength,which makes the power consumption,volume and weight of the system itself large and is limited by the resolution of the detectors.There are many limitations in topographic surveying in complex environments such as shallow waters.The new photon counting detection system lidar greatly improves the detection sensitivity of the system due to the single photon detector and time-correlated single photon counting technology,and can respond to the echo signal of a single photon level,making the system volume,Power consumption and weight are greatly reduced.Therefore,the lidar system based on the photon counting detection system can be mounted on a light and small unmanned aerial vehicle flight platform to achieve rapid acquisition of three-dimensional geographic information of the detection target,and has great advantages and potential in the topographic survey of islands and shallow seas.The airborne radar system needs to go through data processing processes such as denoising,filtering and interpolation to obtain the surveying and mapping products of the island and shallow sea terrain.The denoising of the original data and the filtering of the point cloud are the two most critical steps in data processing.Therefore,in view of the large amount of background noise existing in the original observation data of light and small unmanned aerial vehicle photon counting lidar and the accurate extraction of ground points in the complex terrain environment in the intertidal zone of islands,the relevant research contents mainly include:(1)The research status of photon counting lidar systems,denoising and point cloud filtering algorithms at home and abroad is reviewed,the detection mechanism and noise model of photon counting lidar are studied,and the composition of background noise of photon data of islands,intertidal zones and shallow sea terrain is analyzed.and signal-to-noise ratio characteristics,grasp the complexity of photon counting lidar noise sources and the differences in different terrain environments.(2)A light and small photon counting lidar denoising strategy based on improved local sparsity coefficient(LSC)is proposed.Aiming at the large amount of background noise and signal-to-noise ratio difference in the original observation data obtained by light and small photon counting lidar,firstly,the range of the effective echo signal was determined by the histogram statistics method,and secondly,the grid statistics method was used to preliminarily eliminate the photons far away from the signal.The noise photons in the center are finally denoised by the improved LSC denoising method.The improved LSC method mainly solves two problems in the original LSC method:(1)According to the spatial distribution shape of the signal photons in the grid after coarse denoising,the aggregation degree in the horizontal direction is much larger than that of the noise photons,the circular shape in the original LSC method.The search area cannot accurately identify the noise photons that are closer to the signal photons in the horizontal direction,so an LSC method based on the horizontal elliptical search area is proposed.(2)Aiming at the inaccuracy and poor applicability of the LSC segmentation threshold empirical method in the original LSC method,a method for calculating the LSC segmentation threshold based on the maximum inter-class variance method(OTSU)is proposed to improve the accuracy of segmentation threshold selection.(3)The Jiajing Island area in the South China Sea was selected to carry out the verification experiment of the denoising algorithm.A light and small single-photon lidar UAV flight experiment was carried out,and four different types of photon original observation data were obtained,including mountain vegetation,sandy intertidal zone,nearshore shallower water bodies and deeper water bodies.Using the denoising algorithm proposed in this paper to carry out denoising experiments,and qualitatively and quantitatively analyze the denoising results based on the traditional local distance statistical algorithm and the original LSC algorithm,in the high signal-to-noise ratio island vegetation coverage area and sandy intertidal The average F1-Score of the belt area is 94.6% and 98.9%,and the average F1-Score of the shallow water body area and the deep water body area near the coast with low signal-to-noise ratio are 93.1% and90.7%.The overall F1-Score was 94.3%.The applicability of the improved LSC denoising method proposed in this paper for denoising the original observation data of light and small photon counting lidar is verified.(4)Research the filtering method of laser point cloud data in islands and intertidal zone based on cloth simulation filtering method(CSF).Aiming at the problems of eliminating non-ground points and accurately extracting ground points in point cloud data of islands and intertidal zones,the distribution and morphology of objects in typical experimental areas were analyzed,and the filtering effects of three filtering methods in typical experimental areas were compared.Qualitative and quantitative comparative analysis confirms the reliability of the cloth simulation filtering method in the extraction of ground point clouds in islands and intertidal zones,and provides reliable data support for the subsequent construction of high-precision digital elevation models.