This paper proposes a method based on UAV low-altitude photogrammetry and deep learning algorithms for corn crop growth monitoring. During the shooting process, a unified UAV photogrammetry strategy is set to ensure that the obtained images have high spatial resolution, and after pre-processing the original images, a convolutional neural network (CNN) model is utilized to extract features from the images and improve the accuracy of the CNN with the help of the idea of transfer learning. In addition, multi-scale feature fusion and attention mechanism are introduced to allow the model to focus on important location information, and weighted multi-task loss function is used to jointly optimize the multi-objective values such as plant height, leaf area index, and biomass. Experiments show that the method has good real-time performance and scalability while maintaining high prediction accuracy, providing an effective solution for crop monitoring in precision agriculture.
