Abstract: The radiation duration of lightning signal is random. When the signal length is much less than that of data block, lightning signal will be submerged by background noise using the energy block detection, so the detection probability will be reduced. It is found that the kurtosis can be used to describe the characteristic of the data block of short lightning signal, and moreover, the energy block detection and characteristic detection have complementary properties. Therefore, a real-time adaptive joint detection algorithm based on energy block detection and characteristic detection is proposed to improve the detection probability. And its background noise is estimated adaptively by auto-censoring and excision CFAR technique. The detection performance of this proposed detection algorithm is analyzed and simulated by real lightning data. The simulations and experiments show that this novel joint detection algorithm is effective to improve the detection probability obviously.

Key words: energy block detection, lightning detection, adaptive detection, characteristic detection, joint detection, data block detection

摘要： 雷电电磁辐射的持续时间具有随机性。采用能量法检测雷电数据块，当信号长度远短于数据块长度时，将会产生噪声淹没信号现象而引起检测概率降低的问题。发现可利用峰度来描述含有短雷电信号的数据块的波形特征，而且能量块检测与特征检测具有互补特性。为了提高检测概率，将能量块检测和特征检测相结合，利用自动筛选思想和删余检测技术实时估计背景噪声，提出了实时自适应联合雷电检测算法。通过对实采的雷电数据进行实验，结果表明，所提出的检测算法能够明显提高检测概率，表明了其有效性和实用性。

关键词: 能量块检测, 雷电检测, 自适应检测, 特征检测, 联合检测, 数据块检测