【摘要】 单像素成像作为一种新兴的成像方式，需要通过数学算法重建完整图像。由于成像过程中连续的光信号转化离散信号，量化误差的产生不可避免。因此，探讨量化误差对单像素成像重构效果的影响，并分析不同复杂度待测目标在相同量化深度的重构效果具有重要意义。研究发现，量化误差在单像素成像图像重构过程发挥了重要作用，尤其是在低量化深度的情况下，量化误差对图像质量的影响尤为明显。对于不同复杂度的待测目标，量化误差的影响会导致随着待测目标复杂度的提高，重构图像质量下降。此外，提出了一种分块处理方法，该方法实现了对量化误差空间分布特性的系统性分析，通过将初始图像划分为多个子块，在相同量化深度下对各子块独立进行量化处理，最后将重构子块融合恢复完整图像。有效提高了低量化深度下，单像素成像的重构质量。更多还原

【Abstract】 Single-pixel imaging,as an emerging imaging method,requires reconstructing complete images through mathematical algorithms. Since continuous optical signals are converted into discrete signals during the imaging process,the generation of quantization errors is inevitable. Therefore,it is of great significance to explore the impact of quantization errors on the reconstruction effect of single-pixel imaging and analyze the reconstruction performance of targets with different complexities under the same quantization depth. The study finds that quantization error plays an important role in the image reconstruction process of single-pixel imaging. Especially under the condition of low quantization depth,the impact of quantization error on image quality is particularly significant. For targets to be measured with different complexities,the influence of quantization error leads to a decline in the quality of reconstructed images as the complexity of the target to be measured increases. In addition,a block-based imaging method is proposed. This method enables a systematic analysis of the spatial distribution characteristics of quantization errors. By dividing the original image into multiple sub-blocks,performing independent quantization on each sub-block under the same quantization depth,and finally fusing the reconstructed sub-blocks to recover the complete image,it effectively improves the reconstruction quality of single-pixel imaging at low quantization depths.更多还原