Objective Non-contact infrared imaging thermometry, as an advanced temperature measurement method, holds significant application value in military security, medical health, industrial production, fire detection, and prevention. To address existing issues such as low accuracy and high costs in current infrared thermal imaging, a precise temperature measurement system utilizing MLX90640 infrared array imaging has been developed. This system incorporates an improved least squares curve fitting temperature compensation model and to enhance temperature measurement accuracy. This implementation ensures more efficient system processing and cost-effectiveness.

Methods Collecting object temperature information through the MLX90640 infrared array detector involves gathering data on the temperatures of objects using this technology. The improved bilinear interpolation algorithm is used to expand the temperature data to enhance the display of infrared image detail features. The system temperature calibration and calibration are performed using the BR125 blackbody calibration source for 30 ℃ to 100 ℃, and various temperature test data are collected at different set distances (Tab.1). It is found that the temperature error gradually increases with the increase of the measuring distance, showing a certain linear relationship (Fig.5). Exponential, linear, and second-order polynomial fitting models are established for temperature errors, and the temperature error fitting curve is optimized by comparing the determination coefficients of different temperature error fitting models (Tab.2) to establish a temperature correction optimization model (Tab.3) and estimate the true temperature of the object, achieving accurate prediction of the estimated infrared temperature values at different distances. The uncertainty of the experimental results is analyzed, and uncertainty analysis is carried out in the temperature range of 30 ℃ to 100 ℃. By analyzing the average temperature difference and average standard deviation, the reliability of the experimental results is verified.

Results and Discussions The experimental results show that within the measurement range of 0.2 to 1.2 m, the temperature error after temperature compensation does not exceed 1.0 ℃ (Fig.6). The average temperature error before correction was 5.70 ℃, reduced to 0.24 ℃ after correction, indicating a significant improvement in accuracy post-compensation compared to before. Uncertainty analysis of the temperature compensation model (Tab.4) reveals an experimental average temperature difference of 0.339 ℃ and an average standard deviation of 0.848 ℃ (Tab.5), consistent with expected experimental outcomes, validating the accuracy of the temperature measurement method and the reliability of the proposed model. Prior to system optimization, the edges of the infrared images were blurry; However, after optimization, the imaging system produced more realistic results, with smoother depiction of object edges. Comparison of the infrared image display before and after system improvement shows that the optimized infrared images appear more realistic. The algorithm improves edge smoothness and maintains clarity of object edge details significantly better than the original system. Image curves are more continuous, enhancing overall image quality. By employing enhanced image processing techniques, the algorithm achieves high-resolution images from low-resolution infrared inputs while reducing computational complexity, thereby improving processing speed and reducing resource consumption.

Conclusions Compared to traditional temperature measurement methods, the infrared imaging accurate temperature measurement system has significant advantages, enabling precise non-contact temperature measurement of objects. In contrast to single-point temperature measurement methods, infrared imaging temperature detection provides more comprehensive and rapid measurement of object temperatures, allowing real-time observation of temperature distribution within a specific area. The system design enhances the visual effects of infrared images, while the proposed temperature correction model greatly reduces system temperature measurement errors and conducts uncertainty analysis of experimental results. The thermal imaging temperature measurement system offers high real-time performance and accuracy, providing high-resolution infrared images for precise localization of heat sources. It holds significant application value in fire detection and prevention, security monitoring, and detection of electrical circuit faults causing overheating.