LSS-PR-1.0: Passive Radar Sea and air targets Detection Dataset

The passive radar Sea and air targets detection dataset (LSS-PR-1.0) includes radar range Doppler complex data (RD) and trajectory data (TR) of four types of targets (passenger ships, speedboats, helicopters, drones). The RD data includes 43 target data and 2 sea clutter data; The TR data includes 43 trajectory data, which can support research on the detection, classification, and recognition of typical targets at sea using passive radar.1. Track data (TR data)There are a total of 12 target data items for speedboats, 9 target data items for drones, 11 target data items for helicopters, and 11 target data items for passenger ships.Naming method: Date-PR_TR-Target type and serial number, target batch number Mat (Example: 202306161554-1556_SR_TR_speedboat 01_846.mat, 202306161554-1556 represents the collected RD data time period from 15:54 to 15:56 on June 16, 2023, PR represents passive radar, TR represents trajectory data, speedboat 01 represents the target type as speedboat, and serial number 01846 represents the batch number of the target in the trajectory data).The meaning of each field in each structure of the trajectory data is:Sign --------------- flag, used to determine whether it is a point or a track (point: 0xaaaaaa; track: 0xbbbbbb)Freq - signal frequency (MHz)Update Time - Data update cycle (s)TrgtSize ----------- Target quantityUnixTime ----------- Unix timestamp (ms, the number of milliseconds that have passed since January 1, 1970, 00:00:00 UTC time)Trgt ---------------- Column vector used to store detailed target information (length of trajectory column vector: number of trajectory points * 13)The Trgt field in the trajectory file contains specific information about the target; Every 13 consecutive values of the column vector come from the same trajectory point. From 1 to 13, they represent the track batch number, target longitude, target latitude, target altitude (reserved), target distance, target azimuth, target elevation angle (reserved), target horizontal speed, vertical speed, heading, target echo power, target echo signal-to-noise ratio, and target echo phase of track point 1 in sequence.2. Range Doppler complex data (RD data)There are a total of 12 target data items for speedboats, 9 target data items for drones, 11 target data items for helicopters, 11 target data items for passenger ships, and 2 target data items for sea clutter.Naming method: Start time of collection, Pr-RD_Target type and sequence number Mat (for example: 20230616155443-PRRD_speedboat 01. mat, 20230616 represents the date of data collection, 155443 represents the start time of data collection at 15:54:43, PR represents passive radar, RD represents range Doppler spectral complex data, speedboat 01 represents the target type is speedboat number 01). Each RD data has a TarData structure that contains two variables, RdData and Unit. RdData is a three-dimensional array (for example, TarData.RdData three-dimensional array (5 * 400 * 20) in 20230616155443-PRRD_speedboat 01.mat, where 5 represents taking 5 distance units near the target's location (the calculated distance unit of the target is taken forward by 2 and backward by 2), 400 represents the number of sampling points in the slow time dimension, and 20 represents collecting target data for 20 seconds in one data acquisition); Unit represents the Doppler unit where the target is located, which is calculated based on the position information in the trajectory data.A set of data containing high and low sea states is obtained by taking 100 consecutive distance units of data. The resulting data names are: high sea state (20230828145446Pr-RD_high sea state (10-110). mat), low sea state (20230921144157Pr-RD_low sea state (10-110). mat).For detailed information, please refer to the paper "Chen Xiaolong, Rao Guilin, Key, et al. Passive Radar Low Slow Small Detection Dataset (LSS-PR-1.0) and Multi Domain Feature Extraction and Analysis Method" published in Radar Journal, DOI: 10.12000/JR24145.”

被动雷达低慢小目标检测数据集（LSS-PR-1.0）包含四类目标（客船、快艇、直升机、无人机）的雷达距离多普勒复数据（Range Doppler Complex Data，RD）与轨迹数据（Trajectory Data，TR）。其中RD数据包含43组目标数据与2组海杂波数据；TR数据包含43组轨迹数据，可支撑被动雷达对海上典型目标的检测、分类与识别研究。
1. 轨迹数据（TR数据）
该数据集包含快艇目标数据12组、无人机目标数据9组、直升机目标数据11组、客船目标数据11组。
命名规则：文件命名格式为「采集时间段_PR_TR_目标类型_序号_批次号.mat」，示例：202306161554-1556_SR_TR_speedboat 01_846.mat，其中202306161554-1556代表2023年6月16日15:54至15:56的RD数据采集时间段，PR代表被动雷达，TR代表轨迹数据，speedboat 01代表目标类型为快艇，846为该轨迹数据中目标的批次号。
轨迹数据各结构体字段的含义如下：
Sign：标记位，用于区分点迹与航迹（点迹：0xaaaaaa；航迹：0xbbbbbb）
Freq：信号频率（单位：MHz）
Update Time：数据更新周期（单位：s）
TrgtSize：目标数量
UnixTime：Unix时间戳（单位：ms，即自UTC时间1970年1月1日00:00:00起累计的毫秒数）
Trgt：用于存储目标详细信息的列向量（轨迹列向量长度为轨迹点数×13）
轨迹文件的Trgt字段包含目标的具体信息；该列向量每连续13个数值对应同一轨迹点，按顺序依次为轨迹批次号、目标经度、目标纬度、目标高度（预留字段）、目标距离、目标方位角、目标俯仰角（预留字段）、目标水平速度、垂直速度、航向、目标回波功率、目标回波信噪比、目标回波相位。
2. 距离多普勒复数据（RD数据）
该数据集包含快艇目标数据12组、无人机目标数据9组、直升机目标数据11组、客船目标数据11组，另有海杂波数据2组。
命名规则：文件命名格式为「采集起始时间_PR_RD_目标类型_序号.mat」，示例：20230616155443-PRRD_speedboat 01.mat，其中20230616代表数据采集日期，155443代表数据采集起始时间为15:54:43，PR代表被动雷达，RD代表距离多普勒谱复数据，speedboat 01代表目标类型为01号快艇。
每组RD数据包含TarData结构体，内含RdData与Unit两个变量。RdData为三维数组，例如20230616155443-PRRD_speedboat 01.mat中的TarData.RdData为5×400×20的三维数组，其中5代表以目标位置为中心截取的5个距离单元（目标计算距离单元前后各延伸2个），400代表慢时间维度的采样点数，20代表单次数据采集时长为20秒；Unit代表目标所在的多普勒单元，其计算依据轨迹数据中的位置信息。
通过截取连续100个距离单元的数据，可得到包含高海情与低海情的数据集，对应命名分别为：高海情（20230828145446Pr-RD_high sea state (10-110).mat）、低海情（20230921144157Pr-RD_low sea state (10-110).mat）。
详细信息请参阅发表于《雷达学报》的论文《陈小龙, 饶桂林, Key, 等. 被动雷达低慢小目标检测数据集(LSS-PR-1.0)及多域特征提取与分析方法》，DOI: 10.12000/JR24145。

搜集汇总

背景与挑战

背景概述

LSS-PR-1.0数据集是一个被动雷达低慢小目标检测数据集，包含雷达距离多普勒复数数据（RD）和轨迹数据（TR），涵盖客船、快艇、直升机、无人机四种目标类型以及海杂波数据。该数据集旨在支持被动雷达在海上典型目标的检测、分类和识别研究，数据以结构化格式存储，包括详细的轨迹信息和多普勒频谱数据，适用于雷达信号处理和目标识别领域的实验与分析。