Optimized Deep Neural Network and Its Application in Fine Sowing of Crops

Abstract

Winter wheat is one of the most important food products. Increasing food demand and limited land resources have forced the development of agricultural production to be more refined and efficient. The most important part of agricultural production is sowing. With the promotion of precision agriculture, precision seeding has become the main component of modern agricultural seeding technology system, and the adoption of precision seeding technology is an important means of large-scale production and cost saving and efficiency enhancement. However, the current sowing technology and sowing equipment cannot meet the requirements of wheat sowing accuracy. In this context, a differential perturbation particle swarm optimization (DPPSO) algorithm is proposed by embedding differential perturbation into particle swarm optimization, which shows fast convergence speed and good global performance. After that the DPPSO is used to optimize the convolutional neural network (CNN) to build an optimized CNN (DPPSO-CNN) model and applied to the field of crops fine sowing. Finally, the experimental results show that the proposed method not only has a faster convergence rate but also achieves better wheat seeding performance. The research of this paper an effectively improves the accuracy and uniformity of wheat seeding and lay a foundation for improving wheat yield per unit area and promotes the intelligent development of agriculture in the future.