Using thermal infrared image processing technology of computer vision to protect art and cultural heritage

Abstract

Thermal infrared imaging technology captures the thermal distribution characteristics of object surfaces, offering new possibilities for cultural heritage preservation. These images not only reflect physical conditions but also reveal structural issues, providing crucial technical support for digital conservation. The study outlines acquisition processes including selecting appropriate thermal imaging equipment and controlling environmental factors affecting image quality. It explores preprocessing methods like noise reduction and contrast enhancement to improve usability. By integrating thermal infrared with other data (e.g., visible light), the research demonstrates how multimodal data fusion enhances recognition of surface details and potential damage. For analysis, various feature extraction and classification algorithms—such as Support Vector Machines (SVM) and Neural Networks—are employed to achieve efficient classification and condition assessment. A digital conservation platform was designed and validated through image recognition tests. Results show thermal infrared can effectively identify temperature anomalies on cultural heritage surfaces, which often indicate structural damage. Image fusion techniques enable clearer visualization of heritage details and damage patterns, significantly improving recognition accuracy. The application of feature extraction and classification algorithms further enhances the precision of heritage evaluation. The test results of the digital protection platform show that the platform can effectively manage and display cultural heritage digitally, providing convenience for the protection and research of cultural heritage.