Cloud Imaging Probe (optical array detector with 64 elements) images data

This data set is applied to the manuscript "Classification of Ice Crystal Images Observed from Airborne Cloud Imaging Probes by Transfer Learning". The compressed file contains ice particle data from the Cloud Imaging Probe (optical array detector with 64 elements), the time and location is June 2015 in Beijing, China.(2020-04-29) • The problem of photosensitive edge occlusion and pixel missing of cloud imaging probe particle images is solved by traditional image processing algorithms. • With reference to the classification category of the Cloud Imaging Probe ice crystal pattern parameterization scheme, high-precision ice crystal shape classification is achieved by deep neural networks. • Through the currently popular transfer learning method, TL-ResNet34 is used to complete the nine classifications of ice crystal shapes.