Complex lighting conditions are a critical factor affecting the accuracy of DIC measurements. To obtain high-precision and reliable DIC measurement results, lighting control must be considered a core aspect as important as system calibration and speckle preparation. By carefully designing active uniform illumination, strictly controlling the test environment, optimizing camera settings, ensuring high-quality speckle, and supplementing with appropriate image preprocessing and the selection of robust algorithms, the challenges of light can be overcome to the greatest extent.
Overcoming the effects of light requires a comprehensive approach encompassing lighting design, environmental control, hardware selection, and software algorithms.
1. Optimize active lighting systems (preferred strategy)
Choose a suitable light source: Prioritize high-brightness, stable, and uniformly emitting LED surface or strip light sources. Avoid using light sources that flicker or generate a lot of heat (such as halogen lamps).
Homogenization: Use diffusers, softboxes, light guides, etc. to transform point/line light sources into uniform surface light sources, eliminating directional shadows and highlights.
Multi-light source layout: Multiple light sources are arranged symmetrically to illuminate the object under test from different angles, effectively filling in shadows and improving overall uniformity. Ring light sources are a commonly used choice.
Brightness control: Ensure moderate and adjustable light intensity to avoid underexposure or overexposure. Use the camera histogram or software preview function to assist in adjustment.
Light source stability assurance: Use high-quality regulated power supplies to avoid voltage fluctuations caused by sharing circuits with high-power equipment.
2. Control the test environment
Block ambient light: Conduct tests in a dark room or using a light shield/tent if possible to completely isolate external stray light interference. This is one of the most effective methods.
Stable environment: Avoid areas near switches, other light sources, or direct sunlight.
Camera settings and selection optimization:
Exposure control: Precisely set the exposure time (shutter speed) to ensure appropriate and stable image brightness. Prioritize manual exposure mode to avoid inter-frame brightness fluctuations caused by automatic exposure (AE).
Gain control: Use low gain (or ISO) whenever possible to reduce image sensor noise. In low light conditions, prioritize increasing light source brightness or exposure time rather than increasing gain.
Lens Aperture: Choose an appropriate aperture (F-number) to balance depth of field and light intake. Too small an aperture may cause diffraction and reduce resolution, while too large an aperture will result in a shallow depth of field and may introduce aberrations.
Choose a high-performance camera: Select a camera sensor with high dynamic range (HDR), high quantum efficiency, and low readout noise (such as sCMOS, high-end CMOS), which can tolerate uneven lighting or low light to a certain extent.
3. Optimize speckle quality
High contrast: Ensure sufficient grayscale difference between the speckle (dark) and the background (light). Use a light background with dark speckle when there is insufficient light or a dark background; in strong light, consider a dark background with light speckle.
Appropriate size: The speckle size should match the camera resolution and field of view. It is generally recommended that the speckle diameter be 3-5 pixels (the basis for subpixel algorithms to work).
Randomness and density: The speckle distribution should be highly random and of moderate density (covering most pixels), avoiding regular patterns or large areas of blank/clustered areas.
4. Utilizing advanced image processing and DIC algorithms
Image preprocessing:
Background subtraction/flat field correction: Capture a uniform background image without speckle and subtract it from the original speckle image. This can effectively compensate for fixed illumination inhomogeneities (vignetting) and fixed pattern noise.
Filtering and noise reduction: Before image matching calculation, use appropriate spatial filters (such as Gaussian filtering, median filtering) to reduce image noise, but care should be taken to avoid excessive blurring of speckle edges.
Algorithm Enhancement:
Robust matching criteria: Using matching functions that are relatively insensitive to changes in illumination (such as zero-mean normalized cross-correlation - ZNCC) is more resistant to uneven illumination and slow changes than traditional cross-correlation (CC) or sum of squared differences (SSD).
Research on illumination invariant algorithms: The academic community is exploring more advanced algorithms (such as models based on gradient information, phase information, or deep learning) in an attempt to reduce the impact of illumination changes directly at the algorithm level, but mature industrial applications are still some time away.
Complex lighting conditions are a key factor affecting the accuracy of DIC measurements. By carefully designing active uniform illumination, strictly controlling the test environment, optimizing camera settings, ensuring high-quality speckle patterns, and supplementing with appropriate image preprocessing and robust algorithms, the challenges of light can be overcome to a great extent. Understanding the mechanisms of lighting effects and taking effective countermeasures are crucial to ensuring that DIC technology can unleash its full potential in various demanding real-world applications. In the world of DIC, "seeing" clear and stable speckle patterns is a prerequisite for "measuring" accurate deformation.
