A dataset for dim target detection and tracking of aircraft in radar echo sequences

Abstract

Radar detection and tracking of dim targets under clutter has been a hot yet difficult issue in the fields of military and civilian applications, such as low-altitude security prevention and control, regional situation monitoring, long-range precision strike, battlefield intelligence and reconnaissance. However, there has been problems related to existing radar data, such as poor pertinence of data scenarios, insufficient authenticity of simulation data, nonstandard formatting of measured data, incomplete scene description and calibration information, etc. To meet the data requirements for studies on the radar target detection and tracking technologies under clutter, we took the fixed-wing UAV under clutter as detection object for radar outfield data acquisition and data processing, through which to generate a standard dataset for the radar detection and tracking of dim targets under clutter. This dataset covers typical scenarios such as strong clutter, low signal-to-noise ratio, high dynamics, strong maneuvering, and changing target number. It consists of 15 radar echo sequences in total, and each echo sequence contains a number of pulse signals of a certain temporal duration, a corresponding gate file and a true-value labeling file. With standardized format and accurately labeled information, the dataset can provide basic data for the design and verification of radar dim target detection and tracking algorithms and for studies of clutter characteristics. Data sets in detail: http://www.csdata.org/p/386/.