Abstract: To address issues such as disorganized cigarette cases, product damage and equipment malfunctions caused by vacant spaces formed in the palletizing process of the automated cigarette case truck loading/unloading system, a monitoring system based on computer vision technology was designed. The system consists of two wide-angle cameras, a development board, a light source and an alarm system. The system works as follows: 1) The images captured by the two fixed-mounted wide-angle cameras are stitched together to form an image of large visual field and wide visual angle. 2) Five algorithms (YOLOv5-s, YOLOv7, YOLOv8-s, YOLOv9-s and RT-DETR-R50) are used as candidate models to isolate an individual cigarette case from the background and distinguish between the front row and rear row where the case belongs to. 3) The location of the targeted case is mapped onto a created palletizing space matrix via max-pooling convolution. 4) By dynamic analysis of the palletizing space matrix, vacant space is automatically identified and alarm is triggered real-timely. Application tests of the monitoring system showed that, YOLOv9-s best matched between detection accuracy with speed of the five candidate models. It achieved mAP_0.5 values of 98.3% and 96.4% for the front and rear rows of the palletized cases respectively, and a detection speed of 54 FPS. The vacant space detection accuracy of the monitoring system was 98%, with an average alarm delay of 347 ms, and the frame rate remained above 25 FPS. This system provides technical support for enhancing automation in the warehousing and logistics processes of the tobacco industry.