Get started
Work

Micropulse Lidar Cloud Mask Machine Learning Hand-Labeled Mask

Year: 2026
Type: dataset
Abstract: Cloud detection algorithms of various techniques have been developed and applied to atmospheric ground-based lidar data to identify cloud boundaries and produce clouds masks. While these algorithms are able to identify a wide variety of cloud types and conditions, it is often observed that the algorithms can still fail to accurately detect clouds that are readily discernible when inspecting the lidar imagery. Based on this observation, an alternative approach for cloud detection is to take advantage of machine-learning capabilities and the trained human eye as an interpreter of lidar images, and in turn, to train a neural network to recognize the desired features in the lidar data. This approach has been applied to data from the Atmospheric Radiation Measurement (ARM) user facility’s micropulse lidar (MPL) systems to develop a machine-learning model that can segment images semantically and produce pixel-to-pixel predictions of cloud in lidar images (Cromwell et al. 2019). In addition to these cloud predictions, the model can provide a confidence rating of each pixel prediction. The cloud prediction output from the machine-learning model developed is available in the value-added product (VAP) Micropulse Lidar Cloud Mask Machine-Learning (MPLCMASKML). This VAP provides the cloud mask generated from the prediction output as well as the number of cloud layers and the cloud-layer boundaries. The cloud detection model developed for the MPLCMASKML VAP was trained exclusively with data from the MPL located at the ARM Southern Great Plains (SGP) observatory between January and March of 2015 plus isolated days with heavy aerosol loading. Data from April to July 2016 from the ARM Mobile Facility (AMF3) operated at Oliktok Point (OLI) on the North Slope of Alaska was also used to evaluate the MPLCMASKML VAP.
DOI: 10.5439/17290734
Open Access status: green
OpenAlex ID: W7222655649
Cites:
Cited by:
Related to: