The Dung Beetle Optimization Algorithm (DBO) is characterized by its great convergence accuracy and quick convergence speed. However, like other swarm intelligent optimization algorithms, it also has the disadvantages of having an unbalanced ability to explore the world and to use local resources, as well as being prone to settling into local optimal search in the latter stages of optimization. In order to address these issues, this research suggests a multi-strategy fusion dung beetle optimization method (MSFDBO). To enhance the quality of the first solution, the refractive reverse learning technique expands the algorithm search space in the first stage. The algorithm’s accuracy is increased by adding an adaptive curve to control the dung beetle population size and prevent it from reaching a local optimum. In order to improve and balance local exploitation and global exploration, respectively, a triangle wandering strategy and a fusion subtractive averaging optimizer were later added to Rolling Dung Beetle and Breeding Dung Beetle. Individual beetles will congregate at the current optimal position, which is near the optimal value, during the last optimization stage of the MSFDBO; however, the current optimal value could not be the global optimal value. Thus, to variationally perturb the global optimal solution (so that it leaps out of the local optimal solution in the final optimization stage of the MSFDBO) and to enhance algorithmic performance (generally and specifically, in the effect of optimizing the search), an adaptive Gaussian–Cauchy hybrid variational perturbation factor is introduced. Using the CEC2017 benchmark function, the MSFDBO’s performance is verified by comparing it to seven different intelligence optimization algorithms. The MSFDBO ranks first in terms of average performance. The MSFDBO can lower the labor and production expenses associated with welding beam and reducer design after testing two engineering application challenges. When it comes to lowering manufacturing costs and overall weight, the MSFDBO outperforms other swarm intelligence optimization methods.

Deep learning technology has recently become increasingly prevalent in the field of education due to the rapid growth of artificial intelligence. Teachers’ teaching behavior is a crucial component of classroom teaching activities, and identifying and examining teachers’ classroom teaching behavior is an important way to assess teaching. However, the traditional teaching evaluation method involves evaluating by either listening to the class on-site or playing back the teaching video afterward, which is a time-consuming and inefficient manual method. Therefore, this paper obtained teaching behavior data from a real smart classroom scenario and observed and analyzed the teacher behavior characteristics in this scenario. Aiming at the problems of complex classroom environments and the high similarity between teaching behavior classes, a method to improve YOLOv7 for large target classroom behavior recognition in smart classroom scenarios is proposed. First, we constructed the Teacher Classroom Behavior Data Set (TCBDS), which contains 6660 images covering six types of teaching behaviors: facing the board (to_blackboard, tb), facing the students (to_student, ts), writing on the board (writing, w), teaching while facing the board (black_teach, bt), teaching while facing the students (student_teach, st), and interactive (interact, i). This research adds a large target detection layer to the backbone network so that teachers’ instructional behaviors can be efficiently identified in complex classroom circumstances. Second, the original model’s backbone was extended with an effective multiscale attention module (EMA) to construct cross-scale feature dependencies under various branches. Finally, the bounding box loss function of the original model was replaced with MPDIoU, and a bounding box scaling factor was introduced to propose the Inner_MPDIoU loss function. Experiments were conducted using the TCBDS dataset. The method proposed in this study achieved mAP@.50, mAP@.50:.95, and recall values of 96.2%, 82.5%, and 92.9%, respectively—improvements of 1.1%, 2.0%, and 2.3% over the original model. This method outperformed other mainstream models compared to the current state of the art. The experimental results demonstrate the method’s excellent performance, its ability to identify various classroom behaviors of teachers in realistic scenarios, and its potential to facilitate the analysis and visualization of teacher classroom behaviors.