Taking students as the center,guided by their interest in learning,and aiming to solve problems,we adopt a problem-based teaching method.Taking scene classification based on local binary patterns and its variant features as an example,we design a comprehensive experiment that closely integrates with the teaching syllabus and effectively supports the course objectives.Under the guidance of teachers,students complete the comprehensive experiment in a collaborative manner,connecting the key knowledge of computer vision and image processing course,training the ability to analyze and solve problems,personally experiencing the problem-solving process,and enhancing practice and entrepreneurial innovation abilities.