Segmenting dark-field images of laser-induced damage on large-aperture optics in high-power laser facilities is challenged by complicated damage morphology, uneven illumination and stray light interference. Fully supervised semantic segmentation algorithms have achieved state-of-the-art performance but rely on a large number of pixel-level labels, which are time-consuming and labor-consuming to produce. LayerCAM, an advanced weakly supervised semantic segmentation algorithm, can generate pixel-accurate results using only image-level labels, but its scattered and partially underactivated class activation regions degrade segmentation performance. In this paper, we propose a weakly supervised semantic segmentation method, continuous gradient class activation mapping (CAM) and its nonlinear multiscale fusion (continuous gradient fusion CAM). The method redesigns backpropagating gradients and nonlinearly activates multiscale fused heatmaps to generate more fine-grained class activation maps with an appropriate activation degree for different damage site sizes. Experiments on our dataset show that the proposed method can achieve segmentation performance comparable to that of fully supervised algorithms.

A 1178 J near diffraction limited 527 nm laser is realized in a complete closed-loop adaptive optics (AO) controlled off-axis multi-pass amplification laser system. Generated from a fiber laser and amplified by the pre-amplifier and the main amplifier, a 1053 nm laser beam with the energy of 1900 J is obtained and converted into a 527 nm laser beam by a KDP crystal with 62% conversion efficiency, 1178 J and beam quality of 7.93 times the diffraction limit (DL). By using a complete closed-loop AO configuration, the static and dynamic wavefront distortions of the laser system are measured and compensated. After correction, the diameter of the circle enclosing 80% energy is improved remarkably from 7.93DL to 1.29DL. The focal spot is highly concentrated and the 1178 J, 527 nm near diffraction limited laser is achieved.

Boron nitride (BN) nanosheets incorporated silica antireflective (AR) coating was successfully prepared on fused silica substrate to improve the antilaser-damage ability of transmissive optics used in high-power laser systems. The BN nanosheets were obtained by urea assisted solid exfoliation, and then incorporated into basic-catalyzed silica sols without any further treatment. The transmission electron microscope (TEM) images indicated that the BN nanosheets generally consisted of 2–10 layers. The antireflective BN/$\text{SiO}_{2}$ coating exhibited excellent transmittance as high as 99.89% at 351 nm wavelength on fused silica substrate. The thermal conductivity $0.135~\text{W}\cdot \text{m}^{-1}\cdot \text{K}^{-1}$ of the BN/$\text{SiO}_{2}$ coating with 10% BN addition was about 23% higher than $0.11~\text{W}\cdot \text{m}^{-1}\cdot \text{K}^{-1}$ of the pure $\text{SiO}_{2}$ AR coating. The laser-induced damage threshold (LIDT) of that BN/$\text{SiO}_{2}$ coating is also 23.1% higher than that of pure $\text{SiO}_{2}$ AR coating. This research provides a potential application of BN/$\text{SiO}_{2}$ coatings in high-power laser systems.

SG-III laser facility is now the largest laser driver for inertial confinement fusion research in China. The whole laser facility can deliver 180 kJ energy and 60 TW power ultraviolet laser onto target, with power balance better than 10%. We review the laser system and introduce the SG-III laser performance here.