Research On Systeme And Performence For Shortwave Infrared Laser Non-line-of-sight Three-dimensional Imaging

Laser imaging technology is now widely used in military and civil fields. Among various imaging scenarios, obstacles blocking the line of sight are the most common scenarios. Because conventional imaging methods can not effectively detect the scene out of sight, so some new imaging technology are expected to resolve this problem. Laser non-line of sight(NLOS) imaging is a new imaging technique emerged in recent years which can detect hidden targets by adopting special imaging method similar to corner imaging. With the development of weak light signal detection technology, especially the presence of single photon detector, we can achieve the reconstruction of the hidden scene from the time information using multiple scattering photons counting in the non-line of sight. In view of need of eye-safe applications for non-sight laser imaging, we develop an approach to acquire three-dimensional(3D) reconstruction of multiple NLOS objects with a resolution of approximately 15 cm at shortwave infrared(SWIR) regime.Based on the principles of non-sight imaging systems, we design an imaging system, construct the experimental platform and develop an approach to acquire 3D reconstruction of multiple NLOS objects at eye-safe SWIR regime. The experimental setup consists of a 1550 nm femtosecond fiber with 90 fs pulse width. Because the optical signal undergoes multiple scattering, so its intensity received by the detector is very weak, and reach the photon level. We use the Geiger APD detector with high sensitivity to receive photons whose time resolution is 150 ps. Then we collect data by using the time-correlated single photon counting module. We analysis the various components of the scattered signal,achieve the image reconstruction of multiple targets and accurately restore the shape and the number of hidden target by using back projection reconstruction algorithm.We investigate the three-dimensional reconstruction algorithm and comparative analysis of the two filtering algorithm. The results show that 3D Gaussian-Laplace algorithm has a strong ability to extract the edge information of targets and get image sharpening. We study the influences of filter parameters on the quality of reconstruction, and propose a layered threshold algorithm. We select the layer image with its intensity above 0.7 times than maximum as a threshold for image reconstruction, carry out threshold operation on the Pseudo noise, and effectively eliminate mutual interference between different target signals.We investigate the performance of the non-line sight of laser imaging system, and achieve localization of single point target as it moves toward different directions in the hidden scene. Based on our build system,the vertical and horizontal resolution of dual-point targets can reach 5 and 15 cm, respectively. In the present study, the reconstruction time are investigated and number of different sampling points, range and possible effects of sampling time for different sample points on image quality are analyzed.After the technology is mature, it will be applied for battlefield surveillance, search and rescue, public transportation, medical treatment and other fields.