To maintain public safety and maximize the effectiveness of different events and public areas, crowd management is essential. The crowd is increasing day by day, which creates lots of difficulties for the previous algorithm or somewhere manpower to handle the crowd; to recover the problem, the solution is to use artificial intelligence as it has its vast domain. The Yolov7 algorithm is employed, pretrained on 80 classes, specifically for detecting crowds by identifying individuals. Stocking density is an important factor that affects the health and productivity of animals and the quantity of livestock and poultry produced. However, the present method of counting hemp ducks by hand is laborious, inaccurate, time-consuming, and prone to mistakes like double-counting and omission. Because of this, the aim of this study is to promote the expansion of the smart farming industry by showing how to use deep learning algorithms to keep track of the density of hemp duck flocks in real time.