Research On Key Technologies Of Underwater Photon-counting LIDAR

Photon-counting LIDAR,or single photon LIDAR is a new type of LIDAR which has high sensitivity and time resolution.It incorporates single photon detector which is able to detect as low as a single photon level echo and with the help of time correlated single photon counting(TCSPC)technique,the LIDAR can respond to low intensity signal with high precision.Therefore,photon-counting LIDAR is suitable for echo restricted applications such as remote sensing or low-reflectivity target detection.Currently,photon-counting LIDAR has been widely investigated and applied.For working range,researchers have designed several photon-counting LIDARs which can detect tens to hundreds of kilometers away target.To reduce imaging time,researchers have proposed multibeam,array and frequency multiplex photon-counting LIDARs which can detect many pixels simultaneously.To increase image quality,researchers have developed many reconstruction algorithms based on spatial correlations or statistics difference between signal and noise.However,at present time,little attention has been paid on underwater photon-counting imaging.Compared with traditional LIDAR,applying photon-counting LIDAR on underwater target detection would bring a lot of benefits.Water has a strong attenuation on light and consequently the working range of traditional LIDAR would be restricted.Due to its high sensitivity,photon-counting LIDAR would have much longer working range and photon-counting LIDAR can obtain much clearer three dimension(3D)image for its high time resolution which is very important on many underwater imaging task such as ocean tube inspection.Nevertheless,a few of new problems would arise when photon-counting LIDAR is applied in underwater environment.Water brings a strong backscattering effect on light and the backscattered light would be a significant noise for system.Moreover,backscattered noise is different from background noise.Background noise generally can be considered as constant while backscattered noise varies with range and emitting laser energy.The main task of underwater photon-counting LIDAR is to extract signal counts from abundant noise counts and reconstruct 3D images.To widen the application of photon-counting LIDAR,the thesis investigated several key techniques of underwater photon-counting LIDAR which include properties of light propagation in water,detection probability model of underwater target,noise robust photon-counting imaging method,correction of range walk error for underwater photon-counting LIDAR and low acquisition time photon-counting imaging method.The highlights of these investigations are1)Detection probability model of underwater target.Volume scatter function(VSF)is utilized to quantitatively describe the distribution of backscatter and combining VSF with Poisson detection model,I derived detection probability model of underwater target.Based on the model,the influence of emitting energy and target distance is analyzed.The result revealed that different from land photon-counting LIDAR,higher emitting energy and nearer target distance are not necessary to increase detection probability for underwater photon-counting LIDAR due to the existence of backscatter.It is required to consider both echo level and backscatter level at the same time to acquire high detection probability.2)First signal photon unit photon-counting imaging method for high noise environment.The method can distinguish signal counts from intensive noise counts and it is different from traditional photon-counting imaging methods which generally set fixed acquisition time for every pixel.The method would emit more pulses to low reflectivity pixels while less pulses to high reflectivity pixels,which optimizes the detection process.The simulation and experiment indicated that our method has the more outstanding performance than peak,cross-correlation,unmixing and first photon imaging methods.3)Range walk error(RWE)correction method for underwater photon-counting imaging.Due to the dead time of SPAD,photon-counting LIDAR suffers range walk error problem which is not affordable for high precision imaging.We proposed a range walk error correction method which is based on RWE VS echo photoelectrons(pes)number prior model.We firstly derived the prior relationship between RWE and echo pes number and then we developed a cluster-like algorithm to separate signal and noise counts.With separated noise and signal counts,we estimated the number of echo pes and combining prior model with this echo number,the range walk error can be calculated.4)Photon-counting image with low acquisition time.For some tasks where acquisition time is limited,we proposed a low acquisition time photon-counting imaging method.For hardware,we utilized two SPADs instead of one to accelerate imaging process.For algorithm,a specially-designed reconstruction algorithm is developed which is able to handle the situation where signal count is absent or submerged by noise counts.The reconstruction algorithm is also a utilization of adjacent correlations in images.The land and underwater experiment both revealed that the method can largely reduce acquisition time and thusly it is potential to improve imaging efficiency.We have designed and manufactured a photon-counting LIDAR system which is the basis for our afterwards researches.The purpose of this thesis is to provide exploration on underwater target detection with photon-counting LIDAR.