Large depth-of-field structured light system for 3D microscale measurement with integrated speckle denoising

The increasing demand for high-precision, high-efficiency, and low-cost inspection technologies in industrial manufacturing has driven significant advancements in three-dimensional measurement systems. Structured light measurement, a non-contact and cost-effective three-dimensional measurement method, is widely utilized due to its simplicity and rapid measurement capabilities. However, applying structured light to microscopic scenarios presents challenges, particularly due to limited depth of field caused by the constraints of optical systems. To address this, existing approaches have focused on either physical or virtual extensions of depth of field, each with inherent limitations, such as complex calibration, reduced efficiency, and sensitivity to system parameters. This paper introduces a novel microscopic structured light measurement system with an extended depth of field, leveraging the principle of digital holographic refocusing. Through off-axis holography, light field information from each depth plane is captured, enabling precise reconstruction of defocused fringes. To mitigate the adverse effects of speckle noise from coherent light sources, a residual convolutional neural network-based denoising method is proposed, utilizing simulated datasets for training. Experimental results demonstrate the system’s effectiveness in achieving high accuracy and efficiency in three-dimensional measurements of small samples with rough surfaces as well as for objects exhibiting abrupt height variations such as step-like structures. Additionally, its capability to handle significant material diversity—such as ceramics, matte metals, and low-reflectivity metallic components—highlighting its strong potential for a wide range of industrial applications.