Abstract
Wheat head detection is essential in estimating the important characteristics of wheat. However, detecting wheat heads in images from different domains has been challenging due to variations in domain features and environmental conditions. This research aims to improve the robustness of wheat head detection in wheat images. A combination of Fourier domain adaptation (FDA), adaptive alpha beta gamma correction (AABG) and random guided filter (RGF) preprocessing methods was applied in this study. The authors utilized FDA to reduce variations between different domains by transforming an image into the Fourier domain, aligning its distribution with a randomly selected image of another domain. AABG adjusts image properties based on local statistics of the image patches, and RGF, a technique for edge-aware image filtering, was used as augmentation. An EfficientDet model was trained on the publicly available wheat dataset and the results were analyzed and compared to a baseline model. The FDA + RGF approach achieved an improved mean average precision (mAP) of 0.6534 compared to the baseline mAP of 0.6292. Our study can contribute to advancing wheat head detection techniques in agriculture, addressing factors like variations in wheat head appearance by focusing on improving domain variation through data dependent approaches.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
