| Lidar is a technology which is born from the extension of traditional radar technology to optical band.It has been widely used because of its high ranging accuracy.The single photon detection technology represented by single photon avalanche diode has gradually penetrated into the field of lidar,forming the single photon lidar technology.Single photon avalanche photodiode has the characteristics of high sensitivity and fast response speed,so the detection range of single photon lidar can reach the kilometer level,which makes the long-distance laser 3D imaging become a reality.The existing lidar imaging technology includes single point,line array and planar array imaging methods,etc.And imaging is carried out by means of scanning device.Under the condition of fixed angular resolution,with the increase of target distance,the lateral detail resolution of lidar will be greatly reduced.Using the scanning single photon lidar technology and fusing the sparse three-dimensional point cloud and short-wave infrared image can effectively improve the resolution of long-range targets.Based on the 19-channel single photon detector developed by our laboratory in the past,the linear scanning single photon imaging lidar is designed and implemented,and the post-processing technology of laser point cloud data and the fusion technology of laser point cloud and short wave infrared image are studied.The main contents and contributions are summarized as follows:(1)A line scanning single photon imaging lidar is designed and built.The device adopts the direct measurement of optical time-of-flight method,uses 19 channel InGaAs SPAD single photon detector,and splices into 19 × 1 single photon detection line array through external optical fiber.In order to realize line lighting and detection,a linear laser lighting and receiving light path based on Powell prism is designed.Using one-dimensional galvanometer as scanning component,a galvanometer control software is developed based on the galvanometer software development tool(SDK).Analyzing the synchronization relationship between different signals,and describing the principle and process of the device.(2)The data post-processing technology of single photon lidar is studied.From the working principle of single photon avalanche photodiode,the significance of histogram statistics of echo data is expounded,and the relationship between ranging accuracy and parameters in histogram statistics is analyzed.A signal extraction method based on local threshold extraction is proposed.The method is based on the smooth change characteristics of histogram signal after logarithmic processing,and improves the signal-to-noise ratio by adding window sliding summation.A target echo clustering method is proposed,which is suitable for data with multiple echoes and can separate the distribution of echo data generated by different targets.And the point cloud is denoised based on neighborhood correlation.On the basis of the above,the real-time point cloud processing and display software is realized,and the outdoor experiment is carried out.The experimental results show that the device can clearly image the target about 1 km.(3)The fusion technology of laser point cloud and short-wave infrared image is studied.The principle of calibration and registration between laser point cloud and short-wave infrared image is analyzed.A method to improve the resolution of laser point cloud is proposed,which is based on three-dimensional coordinate interpolation,with the help of the close distance characteristics of adjacent regions,and by fusing short wave infrared image information to complete the laser point cloud.The experimental results show that the method can achieve the effective fusion of 3D point cloud and short-wave infrared image,and obtain the range and reflectivity information of the target at the same time.The fusion results can be used for target segmentation,recognition and other tasks,which will greatly facilitate target detection and recognition. |
PDF Full Text Request