High-precision plant height measurement by drone with RTK-GNSS and single camera for real-time processing

Area of Science:

• Agricultural Engineering
• Remote Sensing
• Computer Vision

Background:

• Conventional drone-based crop height measurement relies on Structure from Motion (SfM) 3D reconstruction, which is computationally intensive and prone to accuracy issues.
• SfM requires multiple aerial images, leading to long processing times and the need for re-capturing images if reconstruction fails.
• Existing methods lack the precision and speed required for real-time agricultural applications.

Purpose of the Study:

• To develop a high-precision, real-time crop height measurement method using a drone equipped with a monocular camera and RTK-GNSS.
• To improve measurement accuracy and reduce computational load compared to traditional SfM techniques.
• To enable rapid, in-flight calibration for variable baseline stereo matching.

Main Methods:

• Utilized a drone with a monocular camera and real-time kinematic global navigation satellite system (RTK-GNSS) for synchronized aerial imaging and positioning.
• Implemented high-precision stereo matching with a long baseline (approx. 1m) by linking RTK-GNSS data with image capture points.
• Developed a novel in-flight calibration method using zero-mean normalized cross-correlation and a two-stage least squares approach for improved accuracy and speed.

Main Results:

• Achieved significant error rate reductions of 62.2% and 69.4% at flight altitudes of 10m and 20m, respectively, compared to conventional methods.
• Demonstrated a depth resolution of 1.6mm with error rate reductions of 44.4% and 63.0% at an altitude of 4.1m.
• The proposed method achieved a processing time of 88ms for high-resolution images (5472x3468 pixels), enabling real-time measurement.

Conclusions:

• The proposed drone-based system offers a highly accurate and efficient solution for real-time crop height measurement.
• The novel in-flight calibration technique enhances stereo matching performance for variable baseline systems.
• This technology has the potential to revolutionize precision agriculture by providing rapid, reliable crop growth data.