Abstract: Due to various environmental factors，the relay platen status recognition needs manual inspection for

a long time，and the current popular platen image recognition program requires a huge background computing

architecture，which can not immediately obtain the verification information in the operation site，and seriously affect the work efficiency of the problem. To address this issue，a solution based on embedded handheld terminal was proposed. The proposed mobile intelligent terminal used a miniature intelligent HD camera to achieve multi-angle shooting of the platen screen cabinet，and effectively solved the problem of hindering shooting due to reflection of glass doors，foreign objects blocking and angles. The proposed relay platen status field recognition method based on embedded and you only look once（YOLO）Nano algorithm selected high performance computing core and special vision recognition module as hardware platform，which avoided huge background computing architecture and realized operation field real-time recognition and verification. The test results show that the system has 100% accuracy rate of relay platen status recognition，which effectively solves the operation site recognition problem of relay platen status and has strong practicality and popularization.

      Key words: relay protection；relay platen status field recognition；embedded image recognition system；

handheld terminals；you only look once（YOLO）Nano algorithm