数字图像相关（Digital Image Correlation, DIC）方法作为一种重要的非接触式全场光学测量技术，在实验力学及相关交叉领域得到了广泛应用。然而，随着研究向多尺度深入，材料在宏观、介观及纳观尺度下呈现出不同的变形机制与特征（如宏观连续变形、介观不连续滑移、纳观晶格畸变等），使得传统DIC方法的适用性与精度面临严峻挑战。针对多尺度变形分析的需求，本文致力于发展和改进适用于不同尺度图像特征的数字图像相关技术。 首先在宏观尺度层面，针对传统DIC方法在测量噪声影响下进行力学参数反演时精度不足的问题，发展了集成式数字图像相关方法和虚场反演法，可以结合弹性力学模型实现残余应力的高精度反演与表征。其次在介观尺度层面，结合介观尺度材料不连续性变形（如滑移、微裂纹等）的测量需求，自主开发了阶跃式DIC测量算法，实现了复杂位移场的测量与表征，突破了传统DIC方法连续性假设的限制。最后在纳观尺度层面，针对高分辨率透射电子显微镜下非均匀变形场的测量需求，建立了一种基于数字图像相关原理的图像处理算法，实现了对透射电镜图像中晶格纳米尺度变形信息的高效、精确提取，并论证了其与传统方法的内在联系与优势。

Digital Image Correlation (DIC), a key non-contact, full-field measurement technique, faces challenges in multi-scale analysis due to varying material deformation mechanisms (macro-continuous, meso-discontinuous, nano-lattice distortion). This thesis develops refined DIC techniques for these distinct scales. At the macro-scale, addressing inversion inaccuracies from noise, an integrated DIC (I-DIC) method and a Virtual Fields Method (VFM) approach were developed, enabling robust, high-precision mechanical parameter identification, including residual stress characterization, using mechanics models. For the meso-scale, addressing discontinuities like slip bands, an improved and applied Heaviside Digital Image Correlation (Heaviside DIC) algorithm accurately captures these features, overcoming conventional DIC's continuity limitations. At the nano-scale, targeting complex High-Resolution Transmission Electron Microscopy (HRTEM) images, a DIC-based automated algorithm was established for precise, efficient lattice nano-deformation extraction, elucidating its connections to and advantages over traditional methods.