AUTOMATED DIY DEPTH FROM FOCUS INSTRUMENT - PLACING SURFACE PROFILOMETRY WITHIN REACH

Mapping surface features has utility across many fields of study, with a variety of methodologies available. LiDAR is used extensively to map surface profiles at the outcrop scale, while atomic force microscopy is geared for mapping surfaces at the nanometer scale. The sizeable gap in scale is currently bridged mainly by photographic methods, which includes the Depth from Focus (DfF) method. DfF involves taking a series of photographs over a range of distances while keeping the focal plane constant. An algorithm then quantifies focus measure for each pixel in each image. With knowledge of the distance between the camera and the sample for each photograph, a surface profile is constructed by finding which distance leads to each pixel being most in focus. DfF instrumentation is commercially available but is generally cost prohibitive for many researchers. We aim to increase the availability of DfF by developing a relatively low-cost DIY alternative. The DfF instrument presented here uses an off-the-shelf linear rail and stepper motor combo, an optical system (i.e., camera or digital microscope), an Arduino, and code developed in MATLAB. For any given iteration, MATLAB signals the camera to take a picture, then drives the linear rail—where the camera is mounted—a set distance forward. Once the image stack is collected, MATLAB proceeds to image registration and focus measure calculations, terminating once the final surface profile is complete. Though this instrumentation is computationally expensive, it needs little oversight and is budget friendly (our setup cost ~$400 for the USB camera setup). Compared to its peers, DfF offers an attractive combination of price (White Light Interferometry: ~$20K, Photogrammetry: ~free) and accuracy (White Light Interferometry: 0.5 microns, Photogrammetry: min. 100 microns). We highlight the capabilities of our DfF instrument by presenting results from several samples at different scales and optical systems. We also discuss sources of error and strategies to decrease these errors. This technique has potential benefit in a variety of fields, including paleontology (e.g., fossil morphology), rheology (e.g., rock deformation, fault analysis), geobiology (e.g., microbial mats), and archaeology (e.g., ancient tool usage).