People Counting Based on YOLO

People counting is an important undertaking with many uses in areas including crowd control, transit, and retail. The final purpose of this project is to create a system that uses video footage to precisely count the number of persons entering and exiting a specified region. The method tracks and detects people in the video using object detection techniques, notably the YOLO (You Only Look Once) algorithm. There are various crucial phases in the project. In order to identify persons in the video frames, the YOLO model is first used. This model has been trained on a sizable dataset and comes with pre-trained weights. For each individual who is detected, the model returns bounding box coordinates, confidence scores, and class labels. The algorithm calculates the number of persons entering and departing the area of interest by comparing the quantity of individuals recognised in the most recent and earlier frames. Based on the change in the count, it is decided whether to go inward or outward. In addition to giving current updates on the population. And their movement status, the created system also combines visual feedback by presenting video frames with bounding boxes around detected individuals. Applications for this data include occupancy management, crowd management, and security surveillance. Experiments with various video datasets serve to confirm the system’s efficacy and accuracy. The results show that even under challenging and busy conditions, the system can reliably measure the number of persons entering and exiting. Overall, this study provides a reliable method for object-based video counting of individuals.