Leaf detection is an important application in the development of both agriculture and ecological research, where biodiversity is more prominent, such as in India. This paper carries out exhaustive survey on existing methods used for leaf detection and compares: YOLOv7, YOLOv8, and Faster R-CNN. These models have different trade-offs regarding computational efficiency versus performance and have been tested for applicability in real-world agricultural applications. While YOLOv7 and YOLOv8 are very efficient, making them suitable for low-resource environments, Faster R-CNN is robust but at the cost of higher computational demands. The results point to the balance between precision and hardware requirements, giving insight into which model could better serve the challenges of leaf detection in agriculture.