Registration of Panoramic/Fish-Eye Image Sequence and LiDAR Points Using Skyline Features

Area of Science:

• Geomatics Engineering
• Computer Vision
• Robotics

Background:

• Accurate registration of LiDAR data and imagery is crucial for Mobile Mapping Systems (MMS).
• Manual registration methods are time-consuming and prone to errors.
• Existing automatic methods may lack precision or robustness.

Purpose of the Study:

• To develop and validate a novel automatic registration method for LiDAR points and panoramic/fish-eye images in MMS.
• To eliminate the need for manual intervention in the registration process.
• To improve the accuracy and efficiency of data fusion in MMS.

Main Methods:

• A rigorous registration model was established between LiDAR points and panoramic/fish-eye images.
• Skyline pixels and LiDAR points were extracted using image pixel values and the registration model.
• A brute-force optimization was employed to match skyline features and refine registration parameters.

Main Results:

• The proposed registration model effectively achieves high-precision alignment of images and LiDAR data.
• The skyline-based method automatically optimizes initial attitude parameters for precise registration.
• Individual attitude correction values are necessary for sequences of images.

Conclusions:

• The developed method provides an effective and automated solution for LiDAR-image registration in MMS.
• Skyline feature matching offers a robust approach for optimizing registration parameters.
• The findings highlight the importance of per-image attitude correction for accurate results.