The use of cranes at construction sites, building zones, development sites, project locations or shipment loading areas, presents a significant threat to workers’ safety, particularly from accidents due to falling loads. This paper is a detailed study that introduces an advanced approach using computer vision techniques, to identify the crane load fall zone and relative location of workers with respect to this load fall zone, thus eliminating future accidents and ensuring worker safety. Previous research studies have faced many challenges in this task especially to accurately identify the crane load, accurately track the crane load, and precisely detect workers’ positions. This paper presents an approach by overcoming these drawbacks by utilizing Segment Anything Model 2.1 (SAM) for crane load tracking and YOLOv11 for worker detection. After the initial target objects are detected in the video-feed (i.e. crane payload and workers), the distance between them is calculated using a centroid-based approach. If the workers are in the proximity of the load being carried, appropriate alerts are sent to the crane operator or site manager. This proposed system greatly increases the safety, reliability and assurance of crane lifting operations.