Samuel Peter Kozaitis

Extended Hough transform for linear feature detection

Improving the accuracy of line segment detection reduces the complexity of subsequent high-level processing common in cartographic feature detection. We developed a new extension to the Hough transform and reported on its application to building extraction. We expanded the Hough space by a third parameter, the horizontal or vertical coordinate of the image space, to provide incremental information as to the length of the lineal feature being sought. Using this extended HT transform allowed us to more accurately detect the true length of a line segment. In addition, we used a Bayesian probabilistic approach to process our extended Hough space that further increased the accuracy of our extended Hough transform.

Multiresolution gradient-based edge detection in noisy images using wavelet domain filters

We detected edges in noisy images using multiresolution analysis with the wavelet transform. products of wavelet coefficients at several scales were used to identify and locate edges. We found that it was important to consider the changes in edge position at different scales to detect edges in noisy imagery. We analyzed one-dimensional edges and compared the results of our approach with the first derivative of the signal. In addition, we compared the results of noisy images with another wavelet-based edge detection method. Our results led to improved edge detection in noisy images.

Multiple-input joint transform correlator for wavelet feature extraction

We describe a joint transform correlator (JTC) that uses multiple input images encoded in the spatial domain for multiwavelet feature extraction. We extend the theory of a JTC to multiple inputs, which enables various combinations of cross-correlations between input images to be performed. Furthermore, we provide experimental results for four inputs with an optically addressed spatial light modulator in the Fourier plane. In addition, it is possible that the space-bandwidth product for multiwavelet feature extraction can be made the same as for a two-input JTC.

Optical wavelet feature extraction using a multiple-input phase-only encoded joint-transform correlator

We describe an optical joint-transform correlator (JTC) that uses phase-only encoded inputs for multi-scale wavelet feature extraction. The analysis of the JTC is extended to multiple phase-only encoded inputs which allowed combinations of cross-correlations between input images to produce a multi-scale wavelet transform under certain conditions. In addition, we compare the results of phase-encoding to that of amplitude-encoding. Using multiple-input phase encoding allows feature extraction to be performed with wavelets described in the spatial domain.

Orbital angular momentum in four channel spatial domain multiplexing system for multi-terabit per second communication architectures

Bandwidth increase has always been an important area of research in communications. A novel multiplexing technique known as Spatial Domain Multiplexing (SDM) has been developed at the Optronics Laboratory of Florida Institute of Technology to increase the bandwidth to T-bits/s range. In this technique, space inside the fiber is used effectively to transmit up to four channels of same wavelength at the same time. Experimental and theoretical analysis shows that these channels follow independent helical paths inside the fiber without interfering with each other. Multiple pigtail laser sources of exactly the same wavelength are used to launch light into a single carrier fiber in a fashion that resulting channels follow independent helical trajectories. These helically propagating light beams form optical vortices inside the fiber and carry their own Orbital Angular Momentum (OAM). The outputs of these beams appear as concentric donut shaped rings when projected on a screen. This endeavor presents the experimental outputs and simulated results for a four channel spatially multiplexed system effectively increasing the system bandwidth by a factor of four.

Wavelet-based noise reduction in multispectral imagery

We used a 3-D wavelet-based denoising method to reduce the noise from multispectral imagery. In our approach, we compared denoising of different bands of a multispectral image using a 2-D denoising technique, by which the wavelet coefficients corresponding to each band were denoised independent of each band, and a 3-D denoising technique by which the wavleet coefficients were denoised by involving all bands in thresholding the wavelet coefficients. Due to the high correlation of the multispectral imagery data along the wavelength axis, the noise can be easily reduced by applying the wavelet transform along the wavelength direction. Our results showed that the 3-D denoising approach improved the overall SNR of a noisy multispectral imagery over the 2-D denoising approach, due to the correlation between the different bands.

Optical correlation using reduced resolution filters

We determined the performance of an optical correlator as the filter SLM was reduced in resolution by a factor of M in a linear dimension. Our configuration allowed a constant reduction, by a factor of M 2 , in the amount of data used to describe the filter when compared to using a filter at full resolution. Reducing the resolution of the filter is shown to primarily affect its impulse response by the addition of copies of the filter image. We provide both general expressions for performance measures and guidelines for minimizing the distortion induced as the resolution of a filter is decreased. Furthermore, we show the effect of reducing the resolution of a binary phase-only filter on autocorrelation experiments and cross-correlation experiments containing competing objects. We obtained experimental results with an optical correlator with a magneto-optic SLM of 128 x 128 pixels in the input plane and a filter of 64 x 64 pixels. The results showed that our arrangement can provide comparable results to a full-resolution filter when the resolution is decreased by a factor of M and the object within the filter images is less than N / M x N / M pixels in size.

Improved feature detection in ECG signals through denoising

We identified important features in ECG signals after using a third-order, correlation-based method for denoising. Using a small sample of test and actual clean ECG signals, we found that the third-order method for denoising preserved the values and location of important peaks better than a conventional second-order wavelet-based method for denoising.

Denoising using higher-order statistics

We used a higher-order correlation-based method for signal denoising. In our approach, we determined which wavelet coefficients contained mostly noise, or signal, based on higher-order statistics. Because the higher that second-order moments of the Gaussian probability function are zero, the third-order correlation coefficient will not have a statistical contribution from Gaussian noise. We obtained results for both 1-D signals and images. In all cases, our approach showed improved results when compared to a more popular denoising method.

Real-time optically processed target recognition system based on arbitrary moire contours

Experiments using Liquid Crystal Televisions (LCTVs) as spatial light modulators for optical correlators, and optical input devices, have been reported upon widely. Moreover, applications of these devices for target recognition and automatic inspection systems are well documented. These systems often require the implementation of computer pre- and post- processing for image filtering and target recognition which handicaps real-time optical processing applications. It is possible to construct custom reference gratings that form a desired moire pattern when mixed with images of structurally illuminated targets. The moire patterns can be in any form, from equal depth contours, to error maps, to any arbitrary pattern desired. We have demonstrated video methods to generate such error maps in real-time. Furthermore, we have removed restrictions on the shape of the output moire contours, thus, developing a real-time automated inspection system based on the optical processing of arbitrary moire contours. We chose the moire pattern to be in the form of a Fresnel zone plate which is sent to an LCTV. Illumination of this zone plate with parallel coherent light results in a diffracted beam which produces a focused line on a detector. The result is a mixed video- optical processing system that could be used for real-time quality level sorting or other automated inspection requirements.