Convolutional neural networks are widely used in tasks such as image recognition due to their excellent performance in two-dimensional data processing. To address the limited AI computing power under the traditional digital integrated circuit computing architecture, this study designs an optical convolution system based on a microlens array. Utilizing the ultra-high-speed and high-parallelism characteristics of optical computing, it achieves convolution computation based on the principles of geometrical optics. The system loads image information and convolution kernels using a spatial light modulator, segments the light field through a microlens array, modulates the segmented sub-regions in parallel using the convolution kernel, and finally converges through the imaging optical path to achieve summation, thus completing the convolution computation in optical space. This study presents the implementation principle of the optical convolution system based on a microlens array, designs the experimental optical path, and proposes implementation methods for key technologies such as system calibration, convolution kernel implementation, and image post-processing. Experimental results show that the system possesses basic imaging performance and the ability to implement complex feature operators such as Gaussian operator, Gradient operator, and Laplacian operator. Limited by device accuracy, the system still has errors caused by aberrations and diffraction. It can provide verification for the feasibility of optical convolution schemes based on geometrical optical architecture in image preprocessing tasks.

This paper presents a refined teaching framework for Tsinghua University's Modern Physics Laboratory course, addressing several common problems in university physics laboratory teaching, including an emphasis on experimental operation over design, results over process, and details over broader scientific perspectives. Guided by national talent-development strategies, Tsinghua University's educational philosophy, and students' future academic and professional needs, the course establishes four core objectives: broadening horizons, building solid foundations, learning to practice, and understanding innovation. To implement these objectives effectively, the teaching team developed a refined and differentiated instructional framework consisting of three major components. First, frontier-oriented experimental chains in condensed matter physics, quantum physics, and optics were established to form a longitudinal framework that reflects academic and technological development. Second, three flexible learning modes were introduced to accommodate students with different backgrounds: a basic mode with six conventional experiments, a guided N+X extended-research mode, and an advanced mode centered on one independent experiment; a process-based assessment system emphasizing students' growth gains was also implemented. Third, a supportive hardware-and-software system was established, including a powerful digital platform, multidimensional communication channels, and a flexible equipment pool for exploratory learning. Two representative teaching cases involving quantum-computing exploration and saturated-absorption spectroscopy design illustrate how students are guided from passive operation toward active inquiry and how an experimental "designer's perspective" is cultivated. Teaching practice indicates that the establishment of these objectives and the implementation of the framework promote students' transition from "operators" to "designers" and help enhance their scientific literacy and innovative capability.