LoRa图传系统在仪表读数识别中的研究

Research on image transmission system by LoRa in instrument recognition

随着人工智能技术的普及应用,在仪表读数识别领域,低成本的解决方案也逐步得到了应用。目前,市面上有基于NB技术对回传图片进行识别的水表读头,但依赖运营商NB网络覆盖和信号质量,只有大城市和县城主要城区才覆盖得较好,在市郊和乡镇农村地区很少有网络覆盖。基于STM32WLE5芯片设计了一套低成本终端采集仪表盘面照片,通过LoRa技术回传图片信息到用户服务平台进行AI识别读数的系统,可以解决NB网络覆盖不到地区的场景应用问题,在市场上具有互补,甚至替代作用。实验结果表明,LoRa图传稳定可靠,满足图像识别的质量要求。

With the widespread application of artificial intelligence technology,low-cost solutions in the field of instrument reading recog-nition have gradually been applied.At present,there are water meter reading heads based on NB technology for image recognition on the market,but they rely on the operator's NB network coverage and signal quality.Only large cities and major urban areas in counties have good coverage,and there is little network coverage in suburban and rural areas.In the paper,a low-cost terminal is designed based on the STM32WLE5 chip to collect dashboard photos.Through LoRa technology,the image information is transmitted back to the user service platform for AI recognition and reading.This system can solve the problem of NB network not covering regions,and has complementary or even substitutive effects in the market.The experimental results show that LoRa image transmission is stable and reliable,meeting the quality requirements of image recognition.