Authors

Abstract

Aiming at the demand of mileage statistics, work area statistics, fault site return and related data automatic retention in the current agricultural machinery reliability appraisal process, the optimization of agricultural machinery video monitoring system based on artificial neural network algorithm was studied. Together with the new video monitoring technology, the agricultural machinery GPS, GSM and fuel consumption recorder technology are combined to realize the functions of real-time data transmission, monitoring, analysis and statistics. Aiming at intelligent fault analysis, a real-time online detection mechanism is proposed, and a cloud collaborative detection mechanism is proposed to solve the problem of inaccurate offline model detection. Use plane map or satellite map to browse. Thus, an online monitoring and visual testing platform for agricultural machinery faults without real-time monitoring records is established. Finally, the test platform is tested and applied. Test results show that the algorithm can greatly shorten the training time and improve the accuracy of training model detection. With the increase of online training iterations, it is helpful to improve the detection accuracy of the generated model. In a word, the system service platform can provide scientific and transparent data for agricultural machinery fault identification, ensure the scientific, open and fair principles of agricultural machinery fault identification, and greatly improve the efficiency of agricultural machinery management.

Abstract in Chinese

针对当前农机可靠性鉴定过程中里程统计、工区统计、故障点返回及相关数据自动保存的需求，研究了基于人工神经网络算法的农机视频监控系统的优化。结合新的视频监控技术，将农机GPS、GSM和油耗记录仪技术相结合，实现实时数据传输、监控、分析、统计等功能。针对智能故障分析，提出了实时在线检测机制，并针对离线模型检测不准确的问题，提出了云协同检测机制。使用平面地图或卫星地图浏览。建立了无实时监测记录的农机故障在线监测与可视化测试平台。最后，对测试平台进行了测试和应用。实验结果表明，该算法可以大大缩短训练时间，提高训练模型检测的准确性。随着在线训练迭代次数的增加，有助于提高生成模型的检测精度。总之，该系统服务平台可以为农机故障识别提供科学、透明的数据，保证农机故障识别的科学、公开、公平原则，大大提高农机管理效率。