YOLOv5-based Object Detection for Food Freezer Warehouses

With the expansion of the scale of the food freezer warehouse, the traditional warehouse management method can no longer adapt to the safety management requirements of frozen food, the detection of warehouse staff and forklifts is an essential element of intelligent management of the food freezer warehouse. The paper uses the YOLOv5 algorithm to implement the detection of staff work clothes and forklifts on a food freezer warehouse dataset created by oneself to comply with the safety management requirements of the food freezer warehouse. The experimental results show that the YOLOv5x-SGD-med model can effectively identify the classes detected in the dataset with P (precision) of 0.985, R (recall) of 0.95, and mAP0.5 of 0.976, which can meet the object detection requirements of the food freezer warehouse.