Passive ranging systems using orthogonal encoding

The hybrid optical-digital system described in this thesis is novel in several ways: (i) the manner in which range information is encoded by the optics, (ii) the signal processing used for decoding the range information, and (iii) the spatially coherent array configurations that were developed and tested. Furthermore, a numerical image simulator tool has been developed and proven valuable in this work. The analytical solutions, simulations, and experimental evidence offer new approaches to signal processing for the orthogonal passive ranging devices.; The ranging system descriptions contained herein include an introduction to the optical mask for range coding, the signal processing theory and algorithms for range decoding, theoretical limitations, and experimental verifications. The optical encoding works by providing an optical magnitude transfer function (MTF, in the spatial-frequency domain) that is essentially a narrow-band filter. The coding is such that the center frequency of this narrow-band filter is a function of range to the object being imaged. Objects in this work are generally assumed to have a broad spatial frequency bandwidth.; The signal processing task, given a range-coded image, is to determine which MTFs, or system modes, generated the data. Fourier transform techniques are used for spectral estimation and peak finding is used for range estimation. Least-squares estimators are used to estimate mode (i.e. range) parameters in a linear statistical model. Signal detectors with constant false-alarm rates are developed as proximity sensors to detect objects over a given set of ranges.; Furthermore, an array of such ranging sensors, when processed in a spatially coherent manner, provides a vast improvement over a single sensor. Each sensor in the ranging array performs an independent range-coding function. However, the array geometry (which also "codes" range information) can be taken into account by summing the shifted images from each element to form an "array image". This image is then processed in a manner similar to the single-aperture image mentioned previously, except that the array responses are included in the system model.