To address the challenges of speckle image blurring and noise coupling caused by thermal airflow disturbances in high-temperature environments, a collaborative restoration method that integrates image degradation theory and multi-frame signal processing is proposed. Firstly, by constructing an adaptive loss function optimization model based on the Structural Similarity Index (SSIM), the degradation parameters are adaptively estimated, breaking through the limitations of the mismatch between the traditional fixed degradation model and the actual thermal disturbances. Then, by combining Wiener filtering and grayscale averaging techniques, the joint optimization of deblurring and denoising is achieved, solving the technical problem of balancing noise suppression and detail preservation. The verification results on a 600 °C high-temperature experimental platform show that the root mean square error of image displacement measurement processed by the traditional grayscale averaging method is 0.006 4 mm; the root mean square error of image displacement measurement processed by the high-temperature Digital Image Correlation (DIC) measurement method based on adaptive optimization of the degradation function and grayscale averaging is 0.004 7 mm, and the image quality is significantly improved, meeting the sub-pixel accuracy requirements. This method does not require the use of complex hardware and does not need to know the parameters of the heat flow field in advance, significantly improving the robustness of high-temperature DIC measurement. It provides a low-cost and high-precision solution for material deformation analysis under extreme working conditions and has important engineering application value.