We test the performances of a Nd:YAG rod with a grooved sidewall in two solar pumped laser setups. In both cases, a Fresnel lens with a surface of $1.03{\mathrm{m}}^2$ is used as the primary solar light concentrator. In the first setup, a ceramic conical cavity is used as the secondary concentrator. Maximum output power of 20.3 W is obtained, corresponding to a slope efficiency of laser output power with respect to focused solar power of 8.34%. In the second setup, a water tube lens is added in a copper conical cavity to further increase the solar energy concentration; from this setup, 27 W output power is obtained, the slope efficiency of laser output power with respect to focused solar power is 9%. In both cases, the performances of the grooved rod are compared with those of an unpolished rod. The efficiency and the beam quality with the grooved rod are superior to those of the unpolished rod.

A free-space optical communication system with a sun light directly pumped laser as the signal transmitter was demonstrated. A 0.6 × 0.6 m Fresnel lens was used as the primary concentrator to collect the solar light. 6.8 W continuous wave laser power was obtained from a 4 mm diameter grooved Nd:YAG rod. The output intensity was modulated with a video signal via a LiNbO ₃ Mach–Zehnder optoelectronic modulator. The video signal with a resolution of 1920 ^* 1080/frame and the frame rate of 25 Hz was transmitted over five-meter free-space in real time with high fidelity. The transmission rate was 125 Mbps and bit error rate was lower than 10 ⁻⁶ . This research shows the feasibility of applying a solar light directly pumped laser for free-space optical communication, which is significant for telecommunications between satellites.

Aiming at the problem of insufficient patient care caused by limited medical resources, a medical human pose dataset is established, a 3D human pose estimation and action recognition algorithm based on lidars is proposed. By adding shallow feature maps, combining prediction branches, and optimizing loss functions, the algorithm reduces the error caused by the occlusion and the small depth difference between human body and background, and improves the pose estimation performance. An action classifier adapted to patient poses is constructed using manually designed features. The recognition accuracy reaches 93.46% and the recognition speed reaches 42FPS on 3080Ti. Experimental results show that the algorithm performs better than other mainstream algorithms on the medical human pose dataset, and can effectively solve the problem of human pose recognition in medical scenes.

Direct generation of vortex beam by mode polarization selection is presented. Due to the dual-polarization property of the laser cavity, two perpendicular polarized transverse modes are simultaneously generated. One is a TEM00 mode and the other is a dark hollow one, which is vortex beam. The topological charge number and the handedness of the generated vortex beam are measured, as well as the spectral and temporal properties. The results show that the beam is a Laguerre-Gaussian mode with the topological charge number of l=±11. Near single frequency mode of the vortex beam is achieved, with a chaotic pulsation temporal output.