Keddy, William Alfred Lyle2024-08-142024-08-1419941994https://hdl.handle.net/1828/18367Two-dimensional (2-0) digital signal processing is applied in many areas including digi­tal image processing, medicine, geography, geology and robotics. In many of these appli­cations 2-D linear, shift invariant (LSI) digital filtering is used and it is desirable to be able to analyze all types of 2-D, LSI filters (finite-duration impulse response (FIR), infi­nite-duration impulse response (IIR) and separable versions of both) and evaluate their performance by applying them to real images. Though many digital image processing (DIP) packages are available, they lack the capability for 2-D filtering beyond simple convolution. The work presented here describes an environment for developing DIP algo­rithms with particular emphasis on the 2-D LSI digital filtering applications. DEDIP, a Developmental Environment for Digital Image Processing, has been developed using the traditional two level programming approach. The low level is a C­-language software library of DIP and system management functions. The high level is a set of digital image processing and file management tools operating under UNIX. Through representation-dependent system management functions, the data and file struc­tures are hidden. These functions perform access and modification, file input and output, and dynamic memory al location on an image which is treated as a single data object. This hides the details of the implementation from the programmer and forms the base of a hierarchical library. The library's top level of DIP functions, built using the system management functions, are data and file structure independent which enhances the portability, reusability and reliability of DEDIP. The DIP algorithms available in DEDIP can be grouped by the operation types of point, intensity and filter. Point operations such as negation and addition of two images are performed on individual pixels. Intensity operations include histogram equalization and linear contrast stretching that modify the histogram of an image. Filter operations include the application of FIR, IIR, separable FIR and separable IIR filter coefficient masks to images as well as the analysis of filter mask characteristics which include the impulse, step, magnitude and phase responses, and the group delays. The capabilities of DEDIP for the implementation of 2-D LSI digital filters are demonstrated through examples of FIR lowpass, IIR lowpass and FIR fan filters, first directional-derivative edge-detection operators and the Laplacian of a Gaussian second­ derivative edge-detection operator. In each example the particular filter coefficient mask is analyzed, the filtering operation on a test image is pe1fo1med and the filtering result is discussed.108 pagesAvailable to the World Wide WebThesis