Object recognition using three-dimensional optical quasi-correlation

A novel method of three-dimensional (3-D) object recognition is proposed. Several projections of a 3-D target are recorded under white-light illumination and fused into a single complex two-dimensional function. After proper filtering, the resulting function is coded into a computer-generated hologram. When this hologram is coherently illuminated, a correlation space is reconstructed such that light peaks indicate the existence and locations of true targets in the observed 3-D scene. Experimental results and comparisons with results of another 3-D object recognition technique are presented.     

Recognition and classification of three-dimensional phase objects by digital Fresnel holography

We demonstrate the validity of wavelet-based processing for recognition and classification of three-dimensional phase objects. One Fresnel digital hologram of each of the three-dimensional (3-D) phase objects to be classified is recorded. The electronic holograms are processed digitally to permit 3-D object information to be retrieved as two-dimensional digital complex images. We use a Mexican-hat wavelet- matched filter (WMF) to enhance the correlation peak and discriminate between the objects. The WMF performs a wavelet transform (WT) to enhance the significant features of the images and the correlation of the WT coefficients thus obtained. We compare the feasibility of a WMF-based object classifier with the matched-filter-based classifier to classify our four 3-D phase objects in a 3-D scene into true or false classes with minimal error.     

Generation of three-dimensional integral images from a holographic pattern of 3-D objects

We present a novel approach for generating three-dimensional (3-D) integral images from a fringe pattern of 3-D objects. A recorded hologram of 3-D objects is segmented into a number of subholograms. Then, different views of 3-D objects are reconstructed from them because each subhologram has its own perspective of 3-D objects in the recording process. These locally reconstructed images can be rearranged as the same subimage array of the conventional integral-imaging system and transformed into virtually picked-up elemental images of 3-D objects. From this newly generated elemental image array, 3-D images could easily be reconstructed by using a white light. Experiments with a 3-D test object have been performed and the results have been presented.     

Three types of computer-generated hologram synthesized from multiple angular viewpoints of a three-dimensional scene

We describe various techniques to synthesize three types of computer-generated hologram (CGH): the Fresnel-Fourier CGH, the Fresnel CGH, and the image CGH. These holograms are synthesized by fusing multiple perspective views of a computer-generated scene. An initial hologram is generated in the computer as a Fourier hologram. Then it can be converted to either a Fresnel or an image hologram by computing the desired wave propagation and imitating an interference process of optical holography. By illuminating the CGH, a 3D image of the objects is constructed. Computer simulations and experimental results underline the performance of the suggested techniques.     

Three-dimensional optical correlator with general complex filters

A new type of electro-optical three-dimensional (3-D) correlator is proposed and demonstrated. A 3-D object scene, observed by multiple cameras from several points of view, is correlated with a 3-D complex computer-generated function. This correlator is a hybridization of the joint transform and the VanderLugt correlators, and, as such, it allows correlations to be made between 3-D real-world objects and 3-D general complex functions. Experimental results are presented.     

Shift- and scale-invariant recognition of contour objects with logarithmic radial harmonic filters

The phase-only logarithmic radial harmonic (LRH) filter has been shown to be suitable for scale-invariant block object recognition. However, an important set of objects is the collection of contour functions that results from a digital edge extraction of the original block objects. These contour functions have a constant width that is independent of the scale of the original object. Therefore, since the energy of the contour objects decreases more slowly with the scale factor than does the energy of the block objects, the phase-only LRH filter has difficulties in the recognition tasks when these contour objects are used. We propose a modified LRH filter that permits the realization of a shift- and scale-invariant optical recognition of contour objects. The modified LRH filter is a complex filter that compensates the energy variation resulting from the scaling of contour objects. Optical results validate the theory and show the utility of the newly proposed method.     

Computer-generated holograms of three-dimensional objects synthesized from their multiple angular viewpoints

Synthesizing computer-generated holograms (CGHs) of a general three-dimensional (3D) object is usually a heavy computational task. We propose and demonstrate a new algorithm for computing CGHs of 3D objects. In our scheme, many different angular projections of computer-designed 3D objects are numerically processed to yield a single two-dimensional complex matrix. This matrix is equivalent to the complex amplitude of a wave front on the rear focal plane of a spherical lens when the object is located near the front focal point and illuminated by a plane wave. Therefore the computed matrix can be used as a CGH after it is encoded to a real positive-valued transparency. When such CGH is illuminated by a plane wave, a 3D real image of the objects is constructed. The number of computer operations are equivalent to those of a two-dimensional Fourier CGH. Computer and optical constructions of 3D objects, both of which show the feasibility of the proposed approach, are described.     

Three-dimensional pattern recognition with a single two-dimensional synthetic reference function

A novel, to our knowledge, method of distortion-invariant three-dimensional (3-D) pattern recognition is proposed. A single two-dimensional synthetic discriminant function is employed as a reference function in the 3-D correlator. Thus the proposed system is able to identify and locate any true-class object in the 3-D scene. Preliminary simulation and experimental results are presented.     

Detecting three-dimensional location and shape of noisy distorted three-dimensional objects with ladar trained optimum nonlinear filters

We propose a filtering technique that uses laser radar (ladar) data to detect a target's three-dimensional (3D) coordinates and shape within an input scene. A two-dimensional ladar range image is converted into 3D space, and then the 3D optimum nonlinear filtering technique is used to detect the 3D coordinates of targets (including the target's distance from the sensor). The 3D optimum nonlinear filter is designed to detect distorted targets (i.e., out-of-plane and in-plane rotations and scale changes) and to be noise robust. The nonlinear filter is derived to minimize the mean of the output energy in response to the input scene in the presence of disjoint background noise and additive noise and to maintain a fixed output peak for the members of the true-class target training set. The system is tested with real ladar imagery in the presence of background clutter. The background clutter used in the system evaluation includes false objects that are similar to the true targets. The correlation output of ladar images shows a dominant peak at the target's 3D coordinates.     

Coordinate-transformed filter for shift-invariant and scale-invariant pattern recognition

A variable radial coordinate transformation of the phase-only filter (POF) that is dependent on the energy's angular distribution of the target spectrum is used to perform shift- and scale-invariant pattern recognition. The POF of a basic size target and the cumulative energy of its angular distribution are calculated. The filter function is then transformed by means of stretching along the radial coordinate so that the same energy contribution to the correlation peak is provided for any size target. The maximum ratio for recognizing scaled objects is 1:1.5. Computer simulations and optical experiments showing the performances of the filter are presented.     

Synthetic discriminant function filter employing nonlinear space-domain preprocessing on bandpass-filtered images

Previously [Appl. Opt. 36, p. 9212 (1997)] we examined the performance of the linear and nonlinear preprocessed difference-of-Gaussians filter, and it was shown that this operation results in greater tolerance to in-class variations while maintaining excellent discrimination ability. The introduction of nonlinearity was shown to provide greater robustness to the filter's response to noise and background clutter in the input scene. We incorporate this new operation into the synthesis of a modified synthetic discriminant function filter. The filter is shown to produce sharp peaks, excellent discrimination without the need to include out-of-class objects, and good invariance to out-of-plane rotation over a distortion range of up to 90 degrees . Additionally, the introduction of nonlinearity is shown to provide greater robustness of the filter response to background clutter in the input scene.     

Effective reduction of the novel look-up table memory size based on a relationship between the pixel pitch and reconstruction distance of a computer-generated hologram

In this paper, we propose an approach, new to our knowledge, to effectively generate and reconstruct the resolution-enhanced computer-generated hologram (CGH) of three-dimensional (3-D) objects with a significantly reduced in memory size novel look-up table (N-LUT) by taking into account a relationship between the pixel pitch and reconstruction distance of the hologram pattern. In the proposed method, a CGH pattern composed of shifted versions of the principal fringe patterns (PFPs) with a short pixel pitch can be reconstructed just by using the CGH generated with a much longer pixel pitch by controlling the hologram reconstruction distance. Accordingly, the corresponding N-LUT memory size required for resolution-enhanced hologram patterns can be significantly reduced in the proposed method. To confirm the feasibility of the proposed method, experiments are carried out and the results are discussed.     

Modified Fresnel computer-generated hologram directly recorded by multiple-viewpoint projections

An efficient method for obtaining modified Fresnel holograms of real existing three-dimensional (3-D) scenes illuminated by incoherent white light is presented. To calculate the hologram, the method uses multiple-viewpoint projections of the 3-D scene. However, contrary to other similar methods, this one is able to calculate the Fresnel hologram of the 3-D scene directly rather than calculating a Fourier hologram first. This significantly decreases the amount of calculations needed to obtain the hologram and also reduces the reconstruction errors. The proposed method is first mathematically introduced and then demonstrated by both simulations and experiments.     

Full in-situ characterization of spatial light modulators in an optical correlator. Filter adaptation to operating curves

Abstract

In this paper we present one method for the characterization of the spatial light modulators (SLMs) of a real-time Vander Lugt type of correlator. This correlator uses two SLMs: one to introduce the scene and a second to introduce the frequency‐matched filter. The SLM characterization methods are in situ, that is in the correlator set‐up. Illumination conditions are different for each SLM, and consequently different characterization techniques must be used in each case. For the characterization of the scene SLM a diffraction method is used, while for the characterization of the filter SLM an interferometric technique is more convenient. Finally, we take into account the operating curves of scene and filter SLMs in the design of the filter in order to optimize the correlation peak.     

Computer-generated holograms of three-dimensional realistic objects recorded without wave interference

We propose a method of synthesizing computer-generated holograms of real-life three-dimensional (3-D) objects. An ordinary digital camera illuminated by incoherent white light records several projections of the 3-D object from different points of view. The recorded data are numerically processed to yield a two-dimensional complex function, which is then encoded as a computer-generated hologram. When this hologram is illuminated by a plane wave, a 3-D real image of the object is reconstructed.     

Scale-invariant optical correlators using ferroelectric liquid-crystal spatial light modulators

New experimental results for scale-invariant implementations of the binary phase-only matched filter and the nonlinear joint transform correlator using ferroelectric liquid-crystal spatial light modulators are presented. We provide a comparative study of both architectures for real-time road-sign recognition. Signal-to-peak-noise ratios in excess of 5 dB over a scale range of 1.0 to 2.0 are achieved under realistic conditions of clutter.     

Implementation of a 3-D Hyperspectral Instrument for Skin Imaging Applications

The many bands acquired by typical multispectral, hyperspectral, and ultraspectral instruments are collected in either a scanning or staring fashion. Staring instruments, like that described in this paper, are popular because they are capable of producing spatially coherent images of a target scene; therefore, they are suitable for noncontact inspection applications. However, the wavebands need to be coregistered. A new application of hyperspectral instrumentation is proposed, where a sheet-of-light method, produced by a low-power laser light, is used to compute range measurements, and a hyperspectral instrument is used to acquire spectral information in the visible (VIS) range of the spectrum. The main advantage of this method is that a single hyperspectral tunable filter arrangement is used to sweep the image to get 3-D information and acquire hyperspectral spectral measurements of a scene. This paper describes the implementation of the image acquisition subsystem, which was done in LabView, and the method used to obtain coregistered radiometrically calibrated spectral imagery and range information. Test cases showed that the instrument may be used to retrieve height and spectral information of small areas and could be used for skin-imaging applications.      

An Internet-enabled image- and model-based virtual machining system

Virtual reality (VR) can be described as a four-dimensional (4-D) simulation of the real world, including the 3-D geometry space, 1-D time and the immersive or semi-immersive interaction interface. VR applications in mechanical-related research areas are becoming popular, e.g. virtual layout design, virtual prototyping, Internet-based virtual manufacturing, etc. However, research in VR applications is facing conflicting requirements for high rendering quality and near real-time interactivity. This paper represents an Internet-based virtual machining system that builds an integrated VR scene, which combines images and models, to overcome the above conflicts. This research is divided into three parts: first, image mosaics techniques are used to implement an Internet-based virtual workshop, which is an image-based virtual scene. The method of obtaining original sequential images, the principle of image mosaics to realize automatic seamless stitching, and projection transformation matrices to reconstruct a closed inward-facing space are presented. Secondly, a model-based virtual milling machine has been constructed with three detailed approaches: a category-based dynamic graph structure to support collision detection, a relation-oriented collision detection method to improve the efficiency of collision detection, and a dynamic modelling method to model a dynamic workpiece object. Finally, an Internet-based virtual milling system, which is the integration of the image-based virtual workshop and the model-based virtual CNC machine, is constructed using the reposition method to achieve visual consistency of the virtual objects and images. This system, which includes an integrated scene, combines the advantages of image-based VR and model-based VR. Consequently, this system has both high rendering quality and good real-time interactivity.     

Holographic femtosecond laser processing using optimal-rotation-angle method with compensation of spatial frequency response of liquid crystal spatial light modulator

Holographic femtosecond laser processing performs high-speed parallel processing using a computer-generated hologram (CGH) displayed on a liquid crystal spatial light modulator. A critical issue is to precisely control the intensities of the diffraction peaks of the CGH. We propose a method of compensating for the spatial frequency response in the design of CGH using the optimal-rotation-angle method. By applying the proposed method, the uniformity of the diffraction peaks was improved. We demonstrate holographic femtosecond laser processing with two-dimensional and three-dimensional parallelism.     

Automated 360° profilometry of a three-dimensional diffuse object and its reconstruction by use of the shading model

The 360° profilometry of a three-dimensional (3-D) diffuse object by use of the light intersection and its image reconstruction by surface shading are presented. The lack of data in one direction, which was due to occlusion, was compensated by the projection of two lines of light from different directions. Some experiments to profile objects and their reconstruction by computer are shown. The entire surface model was constructed, and a real shading image was obtained by means of computer graphics.