Using Mach-Zehnder interferometer with Fourier lens for spectral transformation
Expandable to related experiments, such as 3D display of correlation peaks
Image recognition, optical image encryption
The LBTEK Fourier Transform Image Recognition System is based on Fourier transform theory, utilizing a Mach-Zehnder interferometer paired with a Fourier lens for image spectrum conversion and correlation output. It enables the recognition of simple targets such as identical and different characters, and obtains their joint power spectra and correlation peak images. This experimental setup can be roughly divided into the following three parts: 1) A highly coherent He-Ne laser combined with an attenuator and beam-expanding collimating lens group to provide laser output; 2) A Mach-Zehnder interferometer composed of various mirrors and beam splitters (combiners), with reference images and images to be identified placed separately in the two optical paths of the interferometer, allowing the two images to be projected onto the Fourier transform lens with equal optical path lengths; 3) The use of lenses and a combination of lenses and spatial light modulators to perform two Fourier transforms, obtaining the joint transform power spectrum of the two images to be identified via CMOS1, and finally obtaining the correlation output via CMOS2.
Fourier Transform Image Recognition Systems
Fourier Transform
Mach-Zehnder Interferometer Structure
Image Recognition
The basic principle of the LBTEK Fourier Transform Image Recognition System is: using a laser source with good coherence, the target image to be recognized and the reference image are input in parallel into the coherent optical processing system through a Mach-Zehnder interferometer. The Fourier transform of the lens is utilized to obtain the complex amplitude distribution of the reference image and the target image at the spectral plane. The corresponding power spectrum is formed by CMOS1 and input to the LCD screen of the spatial light modulator. After another Fourier transform by the lens, the correlated output is formed and received by another CMOS2. By observing the brightness and dispersion of the bright spot (correlation peak) formed between the image to be recognized and the reference image, the degree of correlation between them can be determined, thereby identifying the target image. Note: This teaching system is shipped as parts. Users can operate it according to the experimental manual. If installation service is required, the cost will be calculated separately.