Point-cloud-based place recognition using CNN feature extraction
This work addresses place recognition for robotics and autonomous systems, providing an incremental improvement by adapting existing CNN methods to point cloud data with a new dataset for evaluation.
The paper tackles place recognition by proposing a point-cloud-based system that uses a CNN pre-trained on color images to extract features from range images without fine-tuning, achieving significant improvement over hand-crafted features and offering illumination and rotation invariance with robustness against moving objects. It also introduces a new dataset with point cloud and grayscale images covering 360-degree views, organized to facilitate validation of rotation invariance and robustness.
This paper proposes a novel point-cloud-based place recognition system that adopts a deep learning approach for feature extraction. By using a convolutional neural network pre-trained on color images to extract features from a range image without fine-tuning on extra range images, significant improvement has been observed when compared to using hand-crafted features. The resulting system is illumination invariant, rotation invariant and robust against moving objects that are unrelated to the place identity. Apart from the system itself, we also bring to the community a new place recognition dataset containing both point cloud and grayscale images covering a full $360^\circ$ environmental view. In addition, the dataset is organized in such a way that it facilitates experimental validation with respect to rotation invariance or robustness against unrelated moving objects separately.