Optical window freeform surfaces have emerged as a critical research focus in advanced optical engineering owing to their extensive surface degrees of freedom. These surfaces enable the simultaneous correction of on-axis and off-axis aberrations while satisfying stringent requirements for high-performance, lightweight, and compact optical systems. In the initial metrological characterization of these surfaces, coordinate measuring machines (CMMs) are conventionally employed for target point cloud acquisition. However, the achievable measurement accuracy (>2 μm) inherently constrained by CMM precision imposes fundamental limitations for subsequent optical inspections requiring sub-micron to nanometer-level resolution. Meanwhile, although optical measurement methods can result in higher measurement accuracy, they also lead to an increase in costs and testing difficulties. To overcome these limitations, we propose an accelerated point cloud registration methodology based on point-to-triangulation distance estimation. In simulation, using optimal coordinate transformation enabled good capabilities for exceptional surface characterization: peak-to-valley (PV) surface error of 10−6 nm, residual error of 5 nm, and registration accuracy of log10 (mm/°). Further, in the experiment, the PV surface error was reduced from 27.3 μm to 6.9 μm, equivalent to a reduction of 3.95 times. These results confirm that the point-to-triangulation distance approximation maintains sufficient fidelity to the nominal point-to-surface distance, thereby empirically validating the efficacy of our proposed methodology. Notably, compared with conventional 3D alignment methods, our novel 2D estimation registration approach with point-to-triangulation surface normal vectors demonstrates significant advantages in computational complexity, which achieved a 78% reduction from O(n3) to O(n) while maintaining sub-millisecond alignment times. We believe that the method has potential for use as a low-cost optical precision measurement in manufacturing technology.
The uncertainty and confrontational nature of war itself pose significant challenges to the collection and storage of aerial battlefield temporal data. To address the issue of insufficient training of intelligent models caused by the scarcity of air battlefield situation data, this paper designs an air battlefield time series data augmentation model based on a lightweight denoising diffusion probabilistic model (LDMKD-DA). Considering the advantages of a denoising diffusion probabilistic model (DDPM) in processing images, this paper transforms 1D time series data into image data. 1D univariate time series data, such as High-resolution Range Profile dataset, are transformed by Gramian angular fields and Markov transition fields. Multivariate time series data, such as the air target intention dataset, are transformed by matrix expansion. Then, the data augmentation model is constructed based on the denoising diffusion probabilistic model. Considering the need for miniaturization and intelligence in future combat platforms, the depthwise separable convolution is introduced to lighten the DDPM, and, at the same time, the improved knowledge distillation method is introduced to accelerate the sampling process. The experimental results show that LDMKD-DA is capable of generating synthetic data similar to real data with high quality while significantly reducing FLOPs and params, while having significant advantages in univariate and multivariate time series data amplification.