Due to misregistration, update delay, or land cover complexity, they may contain class label noise, i.e., a wrong label assignment.
Charlotte Pelletier, Silvia Valero, Jordi Inglada, Nicolas Champion, Claire MaraisSicre, Gérard Dedieu 2017, 'Effect of Training Class Label Noise on Classification Performances for Land CoverMapping with Satellite Image Time Series', Remote Sensinghttp://www.mdpi.com/2072-4292/9/2/173. Retrieved from DOAJ CC BY 4.0 (https://creativecommons.org/licenses/by-sa/4.0/legalcode)
As a result, the errors in motion estimation can cause misregistration of the resulting point cloud.
Zheng Fang, Shibo Zhao, Shiguang Wen, Yu Zhang 2018, 'A Real-Time 3D Perception and Reconstruction System Based on a 2D Laser Scanner',Journal of Sensorshttp://dx.doi.org/10.1155/2018/2937694. Retrieved from DOAJ CC BY 4.0 (https://creativecommons.org/licenses/by-sa/4.0/legalcode)
Inter- or intra-sensoral misregistration will negatively affect any subsequent image analysis, specifically when processing multi-sensoral or multi-temporal data.
Daniel Scheffler, André Hollstein, Hannes Diedrich, Karl Segl, Patrick Hostert 2017, 'AROSICS: An Automated and Robust Open-Source Image Co-Registration Software for Multi-SensorSatellite Data', Remote Sensinghttps://www.mdpi.com/2072-4292/9/7/676. Retrieved from DOAJ CC BY 4.0 (https://creativecommons.org/licenses/by-sa/4.0/legalcode)