Explore chapters and articles related to this topic
Filters
Published in Afshin Samani, An Introduction to Signal Processing for Non-Engineers, 2019
Filters can also be generalized to be applied to spatial signals, and they can limit the extent of spatial frequency in the filtered image. Similar to frequency-based filters, the spatial filters can also be low-, high- or band-pass filters, but instead of filtering the temporal frequency, they would filter the spatial frequency. Figure 8.9 shows a synthetic image that has two frequency components across the X-axis (i.e., 5 and 60 Hz) and one frequency component across the Y-axis (i.e., 5 Hz). The image is spatially high-pass filtered with a cut-off frequency of 30 Hz across the X- and Y-axes. As can be seen in Figure 8.9, low-frequency components across the X- and Y-axes are very clear, meaning that if one looks at the horizontal and vertical directions of the image, dark and bright bands across the X, Y axis is quite clear. However, this pattern is quite attenuated in the filtered version of the image. Please note I do not mean the narrow bright and dark bands but the wider bands which are visible in the original image but not in the filtered image.
Digital Image Manipulations
Published in Julio Sanchez, Maria P. Canton, William Perrizo, Space Image Processing, 2018
Julio Sanchez, Maria P. Canton, William Perrizo
Spatial filters are implemented by means of a process called spatial convolution or finite impulse response (FIR) filtering. In this case, a two-dimensional pixel kernel is moved across the image, pixel by pixel. The result of a mathematical operation on the elements in the kernel is placed in the output set. The calculation is defined as a linear process since each of the elements in the kernel is multiplied by a constant factor called the convolution coefficient. The convolution coefficient can be expressed as
Glossary of Computer Vision Terms
Published in Edward R. Dougherty, Digital Image Processing Methods, 2020
Robert M. Haralick, Linda G. Shapiro
A spatial filter is an image operator in which the spatial and transform domain have similar geometries and in which the image output value at each pixel depends on more than one pixel value in the input image. Usually, but not always, the image output value has its highest dependence on the image input values in some neighborhood centered in the corresponding pixel in the input image.
Deep convolutional neural network for three-dimensional objects classification using off-axis digital Fresnel holography
Published in Journal of Modern Optics, 2022
B. Lokesh Reddy, R N Uma Mahesh, Anith Nelleri
Digital holography is an interferometric imaging technique, a single-shot off-axis digital hologram recording technique that is well suitable for imaging 3D objects. Figure 2 shows the experimental set-up of an off-axis digital holographic system configured in transmission mode using Mach–Zehnder interferometer geometry. The light beam emitted from the laser source (He–Ne laser) of wavelength is filtered using a spatial filter and collimated with the collimation lens of focal length . The collimated beam enters the Mach–Zehnder interferometer by splitting the beam into two arms using a BS1. In the object arm, a 3D object is placed at a distance ‘’ from the recording plane to form the object wave, and another arm forms the reference wave. At the interferometer exit (BS2), the object and reference waves are interfered with at an angle to form an off-axis digital Fresnel hologram. The CMOS sensor with square pixel pitch is used to record the off-axis digital Fresnel hologram.
Multi view interferometric tomography measurements of convective phenomena in a differentially-heated nanofluid layer
Published in Experimental Heat Transfer, 2022
S. Srinivas Rao, Atul Srivastava
The multi-view projection data of the convective flow field in the cylindrical tomographic test cell for any given working fluid medium have been mapped using the Mach-Zehnder interferometer. The assembly of the optical components for development of the laser interferometry was schematically shown in Figure 3. A He-Ne laser of 632.8 nm wavelength of coherent light source is aligned in-line with the spatial filter (objective lens of 40×) to form a diverging source of light beam. The diverging light beam was collimated using an achromatic doublet lens, through placing the lens at its focal length from the spatial filter. The optical components consigned for assembling the present interferometer are consisting of 100 mm diameter size lens intended for achromatic doublet, mirrors, and beam splitters. The collimator consents to form an appropriate working collimated light beam of 60 mm in diameter. The collimated light beam is passed through first beam splitter (BS1) that splits light beam into two with 50% reflectivity and 50% transmissivity. The transmitted light beam is used as reference arm of the interferometer, whereas the reflected light beam from BS1 is used as the test arm of the interferometer by placing the test cavity in the direction of the light beam.
Vertically Polarized Laser Speckle Contrast Imaging to Monitor Blood Flow in Pulp
Published in Journal of Modern Optics, 2021
Fenfen Xu, Chengfeng Xie, Yubao Zhang, Gang Shi, Jiulin Shi, Xiaojun Xu, Youjiang Zhao, Yirui Zhu, Xingdao He
As shown in Figure 1, the vp-LSCI system was composed of a randomly polarized He–Ne laser (22.5 mW, HNL225R, Thorlabs, USA) with a wavelength of 632.8 nm, a polarizer, a 40× space filter, a 10 mm convex lens, an analyzer, a 10-bit CMOS camera (480 × 640 resolution, MER-031-860U3M/C-L, Daheng Imaging, China), and an imaging lens. In the system, the direction of the polarizer used in the lighting system was perpendicular to that of the analyzer used in the imaging system. The light emitted by a random polarized He–Ne laser was transformed into a linearly polarized light through a polarizer, and then the beam was expanded and collimated by a spatial filter and a convex lens device, and then it was uniformly irradiated on the tooth. The scattered light passing through the tooth was adjusted to vertical polarization by the analyzer and collected speckle images by CMOS with an imaging lens. The aperture of the imaging lens was adjusted so that the minimum resolvable speckle size is at least two pixels in size to meet the Nyquist sampling law.