Timely and accurate agricultural monitoring can provide effective guidance for crop management but the process can be challenging. In this study, a GF-2 remote sensing image was used as the research object to extract the winter wheat based on an improved deforming full convolutional neural network (DFCN) model. The deformable convolution model was developed to try to better extract geometric features and object-level semantic information and improve the ability to describe the different spatial distribution characteristics of winter wheat. Convolution deformation was obtained by adding a trainable two-dimensional migration to each convolution layer in the network. The DFCN network model was compared to traditional neural networks such as FCN and U-Net. The precision ratio of DFCN was improved to 98.1%,and the time required was reduced to 0.63 seconds. In contrast, the accuracies of winter wheat automatic interpretation obtained by FCN and U-Net were 89.3% and 93.9%, respectively. The experimental results show that using the DFCN model to interpret winter wheat automatically can achieve the highest accuracy, better capturing its morphology and spatial distribution, and with good generalization performance. Through the training of the input data set and parameter tuning, a high-precision intelligent and DFCN model with high practicability and reliability was constructed, and a high-precision automatic remote sensing identification system for typical crops in the north of the Huang-Huai-Hai Plain was realized.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.