To address the contradiction between the high-concurrency processing and low-latency response of measurement data generated by photoelectric sensing devices，this paper proposes a distributed computing architecture oriented toward photoelectric real-time measurement. By analyzing the characteristics of Flink and Spark Streaming，a scheduling model based on task priority and adaptive resource load is constructed to achieve dynamic resource management for streaming data. The model attains an event-detection accuracy of 96. 92%， a peak throughput of 132. 5 GB/min， protocol processing capability of 3 492 packets/s，and a 31. 7% reduction in computing resource overhead. Test results show that the fault recovery time is reduced to 153. 1 ms(minimum 116 ms)while data-processing integrity reaches 98. 77%(peak 99. 9%). During 72 h of continuous operation，the model maintained a throughput of 124. 7 GB/min and a detection accuracy of 94. 8%，with a comprehensive resource-optimization rate of 31. 7%(20. 9% and 48. 1% reductions in computational complexity and memory overhead，respectively). These results verify its robustness and elastic optimization capability under complex multivoltage levels，thus providing effective reference for establishing real-time analysis platforms in power systems based on photoelectric sensing.