Wang Xiaodong, Han Shuaifeng
College of Computer and Information Engineering, Henan Normal University, Xinxiang 453007, China.
Sensors (Basel). 2024 Sep 17;24(18):6003. doi: 10.3390/s24186003.
Multi-spectral temperature measurement technology has been found to have extensive applications in engineering practice. Addressing the challenges posed by unknown emissivity in multi-spectral temperature measurement data processing, this paper adds emissivity constraints to the objective function. It proposes a multi-spectral radiation temperature measurement data processing model realized through a particle swarm optimization algorithm improved based on multiple strategies. This paper simulates six material models with distinct emissivity trends. The simulation results indicate that the algorithm calculates an average relative temperature error of less than 0.3%, with an average computation time of merely 0.24 s. When applied to the temperature testing of silicon carbide and tungsten, experimental data further confirmed its accuracy: the absolute temperature error for silicon carbide (tungsten) is less than 4 K (7 K), and the average relative error is below 0.4% (0.3%), while two materials maintain an average computation time of 0.33 s. In summary, the improved particle swarm optimization algorithm demonstrates strong performance and high accuracy in multi-spectral radiation thermometry, making it a feasible solution for addressing multi-spectral temperature measurement challenges in practical engineering applications. Additionally, it can be extended to other multi-spectral systems.
多光谱温度测量技术在工程实践中具有广泛的应用。针对多光谱温度测量数据处理中发射率未知所带来的挑战，本文在目标函数中加入发射率约束。提出了一种基于多种策略改进的粒子群优化算法实现的多光谱辐射温度测量数据处理模型。本文模拟了六种具有不同发射率趋势的材料模型。模拟结果表明，该算法计算出的平均相对温度误差小于0.3%，平均计算时间仅为0.24 s。将其应用于碳化硅和钨的温度测试时，实验数据进一步证实了其准确性：碳化硅（钨）的绝对温度误差小于4 K（7 K），平均相对误差低于0.4%（0.3%），且两种材料的平均计算时间均为0.33 s。综上所述，改进后的粒子群优化算法在多光谱辐射测温中表现出强大的性能和较高的精度，使其成为解决实际工程应用中多光谱温度测量挑战的可行方案。此外，它还可扩展到其他多光谱系统。
