【摘要】 基于人工智能震相识别的实时智能地震处理系统(简称“RISP”系统)，近年来在国内地震的余震处理中展现出优异的自动编目能力。为评估“RISP”系统在黑龙江测震台网实时处理能力和地震检测精确程度，通过对2024年1—3月波形数据进行实时在线检测，并对2017—2020年历史波形数据进行自动编目。将自动检测结果与人工编目结果对比分析发现：天然地震的匹配率为97.04%，非天然地震的的匹配率为80.13%；实时在线监测的准确率为76.14%，离线波形回溯的准确率为91.97%。结果表明，“RISP”系统在黑龙江测震台网产出的自动编目与人工编目具有较好的一致性，且对非天然地震事件处理准确、可靠，具备应用于矿山、水库等专业监测台网的潜力。更多还原

【Abstract】 The real-time intelligent seismic processor (RISP) is an AI-based system that has demonstrated excellent automatic cataloging capabilities in recent domestic earthquake aftershock processing.To evaluate the real-time processing capabilities and earthquake detection accuracy of the “RISP” system in the Heilongjiang Seismic Network,this paper performs real-time online detection on waveform data from January to March 2024 and automatic cataloging on historical waveform data from 2017 to 2020.By comparing the automatic detection results with manual cataloging results,we find that the match rate for natural earthquakes is 97.04%,and the match rate for non-natural earthquakes is 80.13%;the accuracy rate of real-time online monitoring is 76.14%,and the accuracy rate of offline waveform backtracking is 91.97%.Therefore,the automatic catalogingresult produced by the “RISP” system in the Heilongjiang Seismic Network shows good consistency with manual cataloging result,and the system is accurate and reliable in processing non-natural earthquake events,demonstrating potential for application in specialized monitoring networks such as mines and reservoirs.更多还原