Much effort has been spent over the last decade in producing so called "Machine Vision" systems for use in robotics, automated inspection, assembly and numerous other fields. Because of the large amount of data involved in an image (typically ¼ MByte) and the complexity of many algorithms used, the processing times required have been far in excess of real time on a VAX-class serial processor. We review a number of image understanding algorithms that compute a globally defined "state", and show that they may be computed using simple local operations that are suited to parallel implementation. In recent years, many massively parallel machines have been designed to apply local operations rapidly across an image. We review several vision machines. We develop an algebraic analysis of the performance of a vision machine and show that, contrary to the commonly-held belief, the time taken to relay images between serial streams can exceed by far the time spent processing. We proceed to investigate the roles that a variety of pipelining techniques might play. We then present three pipelined designs for vision, one of which has been built. This is a parallel pipelined bit slice convolution processor, capable of operating at video rates. This design is examined in detail, and its performance analysed in relation to the theoretical framework of the preceeding chapters. The construction and debugging of the device, which is now operational in its hardware is detailed.

This study was concerned with the computer automation of land evaluation. This is a broad subject with many issues to be resolved, so the study concentrated on three key problems: knowledge based programming; the integration of spatial information from remote sensing and other sources; and the inclusion of socio-economic information into the land evaluation analysis. Land evaluation and land use planning were considered in the context of overseas projects in the developing world. Knowledge based systems were found to provide significant advantages over conventional programming techniques for some aspects of the land evaluation process. Declarative languages, in particular Prolog, were ideally suited to integration of social information which changes with every situation. Rule-based expert system shells were also found to be suitable for this role, including knowledge acquisition at the interview stage. All the expert system shells examined suffered from very limited constraints to problem size, but new products now overcome this. Inductive expert system shells were useful as a guide to knowledge gaps and possible relationships, but the number of examples required was unrealistic for typical land use planning situations. The accuracy of classified satellite imagery was significantly enhanced by integrating spatial information on soil distribution for Thailand data. Estimates of the rice producing area were substantially improved (30% change in area) by the addition of soil information. Image processing work on Mozambique showed that satellite remote sensing was a useful tool in stratifying vegetation cover at provincial level to identify key development areas, but its full utility could not be realised on typical planning projects, without treatment as part of a complete spatial information system.

The aims of the project were twofold: 1) To investigate classification procedures for remotely sensed digital data, in order to develop modifications to existing algorithms and propose novel classification procedures; and 2) To investigate and develop algorithms for contextual enhancement of classified imagery in order to increase classification accuracy. The following classifiers were examined: box, decision tree, minimum distance, maximum likelihood. In addition to these the following algorithms were developed during the course of the research: deviant distance, look up table and an automated decision tree classifier using expert systems technology. Clustering techniques for unsupervised classification were also investigated. Contextual enhancements investigated were: mode filters, small area replacement and Wharton's CONAN algorithm. Additionally methods for noise and edge based declassification and contextual reclassification, non-probabilitic relaxation and relaxation based on Markov chain theory were developed. The advantages of per-field classifiers and Geographical Information Systems were investigated. The conclusions presented suggest suitable combinations of classifier and contextual enhancement, given user accuracy requirements and time constraints. These were then tested for validity using a different data set. A brief examination of the utility of the recommended contextual algorithms for reducing the effects of data noise was also carried out.

The northern half of the parish of st. catherine .in Jamaica was selected as a test area to study 1 by rreans of rerrote sensing, the problerrs of soil erosion in a tropical environment. An initial study was carried out to determine whether eroded land within this environment could be successfully interpreted and mapped from the available 1: 25, a::xJ scale aerial photographs. When satisfied that a sufficiently high percentage of the eroded land could be interpreted on the aerial photographs the main study was initiated. This involved interpreting the air photo cover of the study area for identifying and classifying land use and eroded land, and plotting t..lJ.e results on overlays on tOfX)graphic base maps. These overlays were then canposited with data on the soils and slopes of the study area. The areas of different soil t"tpe/slope/land use canbinations were then rreasured, as was the area of eroded land for each of these combinations. This data was then analysed in two ways. The first way involved deter:mining which of the combinations of soil type, slope and land use were most a.'1d least eroded and, on the basis of this, to draw up recomrendations concerning future land use. The second analysis was aimed at deter:mining which of the three factors, soil type, slope and land use, was rrost responsible for dete:rmining the rate of erosion. Although it was possible to show that slope was not very significant in detennining the rate of erosion, it was nruch rrore difficult to separate the effects of land use and soil type. The results do, however, suggest that land use is rrore significant tr.an soil type in detennirling the rate of erosion within the study area.

The ERS-1 Satellite was launched in July 1991 by the European Space Agency into a polar orbit at about 800 km, carrying a C-band scatterometer. A scatterometer measures the amount of backscatter microwave radiation reflected by small ripples on the ocean surface induced by sea-surface winds, and so provides instantaneous snap-shots of wind flow over large areas of the ocean surface, known as wind fields. Inherent in the physics of the observation process is an ambiguity in wind direction; the scatterometer cannot distinguish if the wind is blowing toward or away from the sensor device. This ambiguity implies that there is a one-to-many mapping between scatterometer data and wind direction. Current operational methods for wind field retrieval are based on the retrieval of wind vectors from satellite scatterometer data, followed by a disambiguation and filtering process that is reliant on numerical weather prediction models. The wind vectors are retrieved by the local inversion of a forward model, mapping scatterometer observations to wind vectors, and minimising a cost function in scatterometer measurement space. This thesis applies a pragmatic Bayesian solution to the problem. The likelihood is a combination of conditional probability distributions for the local wind vectors given the scatterometer data. The prior distribution is a vector Gaussian process that provides the geophysical consistency for the wind field. The wind vectors are retrieved directly from the scatterometer data by using mixture density networks, a principled method to model multi-modal conditional probability density functions. The complexity of the mapping and the structure of the conditional probability density function are investigated. A hybrid mixture density network, that incorporates the knowledge that the conditional probability distribution of the observation process is predominantly bi-modal, is developed. The optimal model, which generalises across a swathe of scatterometer readings, is better on key performance measures than the current operational model. Wind field retrieval is approached from three perspectives. The first is a non-autonomous method that confirms the validity of the model by retrieving the correct wind field 99% of the time from a test set of 575 wind fields. The second technique takes the maximum a posteriori probability wind field retrieved from the posterior distribution as the prediction. For the third technique, Markov Chain Monte Carlo (MCMC) techniques were employed to estimate the mass associated with significant modes of the posterior distribution, and make predictions based on the mode with the greatest mass associated with it. General methods for sampling from multi-modal distributions were benchmarked against a specific MCMC transition kernel designed for this problem. It was shown that the general methods were unsuitable for this application due to computational expense. On a test set of 100 wind fields the MAP estimate correctly retrieved 72 wind fields, whilst the sampling method correctly retrieved 73 wind fields.

SPOT simulation imagery was acquired for a test site in the Forest of Dean in Gloucestershire, U.K. This data was qualitatively and quantitatively evaluated for its potential application in forest resource mapping and management. A variety of techniques are described for enhancing the image with the aim of providing species level discrimination within the forest. Visual interpretation of the imagery was more successful than automated classification. The heterogeneity within the forest classes, and in particular between the forest and urban class, resulted in poor discrimination using traditional `per-pixel' automated methods of classification. Different means of assessing classification accuracy are proposed. Two techniques for measuring textural variation were investigated in an attempt to improve classification accuracy. The first of these, a sequential segmentation method, was found to be beneficial. The second, a parallel segmentation method, resulted in little improvement though this may be related to a combination of resolution in size of the texture extraction area. The effect on classification accuracy of combining the SPOT simulation imagery with other data types is investigated. A grid cell encoding technique was selected as most appropriate for storing digitised topographic (elevation, slope) and ground truth data. Topographic data were shown to improve species-level classification, though with sixteen classes overall accuracies were consistently below 50%. Neither sub-division into age groups or the incorporation of principal components and a band ratio significantly improved classification accuracy. It is concluded that SPOT imagery will not permit species level classification within forested areas as diverse as the Forest of Dean. The imagery will be most useful as part of a multi-stage sampling scheme. The use of texture analysis is highly recommended for extracting maximum information content from the data. Incorporation of the imagery into a GIS will both aid discrimination and provide a useful management tool.

Textured regions in images can be defined as those regions containing a signal which has some measure of randomness. This thesis is concerned with the description of homogeneous texture in terms of a signal model and to develop a means of spatially separating regions of differing texture. A signal model is presented which is based on the assumption that a large class of textures can adequately be represented by their Fourier amplitude spectra only, with the phase spectra modelled by a random process. It is shown that, under mild restrictions, the above model leads to a stationary random process. Results indicate that this assumption is valid for those textures lacking significant local structure. A texture segmentation scheme is described which separates textured regions based on the assumption that each texture has a different distribution of signal energy within its amplitude spectrum. A set of bandpass quadrature filters are applied to the original signal and the envelope of the output of each filter taken. The filters are designed to have maximum mutual energy concentration in both the spatial and spatial frequency domains thus providing high spatial and class resolutions. The outputs of these filters are processed using a multi-resolution classifier which applies a clustering algorithm on the data at a low spatial resolution and then performs a boundary estimation operation in which processing is carried out over a range of spatial resolutions. Results demonstrate a high performance, in terms of the classification error, for a range of synthetic and natural textures.

Image segmentation is one of the most computationally intensive operations in image processing and computer vision. This is because a large volume of data is involved and many different features have to be extracted from the image data. This thesis is concerned with the investigation of practical issues related to the implementation of several classes of image segmentation algorithms on parallel architectures. The Transputer is used as the basic building block of hardware architectures and Occam is used as the programming language. The segmentation methods chosen for implementation are convolution, for edge-based segmentation; the Split and Merge algorithm for segmenting non-textured regions; and the Granlund method for segmentation of textured images. Three different convolution methods have been implemented. The direct method of convolution, carried out in the spatial domain, uses the array architecture. The other two methods, based on convolution in the frequency domain, require the use of the two-dimensional Fourier transform. Parallel implementations of two different Fast Fourier Transform algorithms have been developed, incorporating original solutions. For the Row-Column method the array architecture has been adopted, and for the Vector-Radix method, the pyramid architecture. The texture segmentation algorithm, for which a system-level design is given, demonstrates a further application of the Vector-Radix Fourier transform. A novel concurrent version of the quad-tree based Split and Merge algorithm has been implemented on the pyramid architecture. The performance of the developed parallel implementations is analysed. Many of the obtained speed-up and efficiency measures show values close to their respective theoretical maxima. Where appropriate comparisons are drawn between different implementations. The thesis concludes with comments on general issues related to the use of the Transputer system as a development tool for image processing applications; and on the issues related to the engineering of concurrent image processing applications.

Digital image processing is exploited in many diverse applications but the size of digital images places excessive demands on current storage and transmission technology. Image data compression is required to permit further use of digital image processing. Conventional image compression techniques based on statistical analysis have reached a saturation level so it is necessary to explore more radical methods. This thesis is concerned with novel methods, based on the use of fractals, for achieving significant compression of image data within reasonable processing time without introducing excessive distortion. Images are modelled as fractal data and this model is exploited directly by compression schemes. The validity of this is demonstrated by showing that the fractal complexity measure of fractal dimension is an excellent predictor of image compressibility. A method of fractal waveform coding is developed which has low computational demands and performs better than conventional waveform coding methods such as PCM and DPCM. Fractal techniques based on the use of space-filling curves are developed as a mechanism for hierarchical application of conventional techniques. Two particular applications are highlighted: the re-ordering of data during image scanning and the mapping of multi-dimensional data to one dimension. It is shown that there are many possible space-filling curves which may be used to scan images and that selection of an optimum curve leads to significantly improved data compression. The multi-dimensional mapping property of space-filling curves is used to speed up substantially the lookup process in vector quantisation.

This thesis describes the development of an operational river basin water resources information management system. The river or drainage basin is the fundamental unit of the system; in both the modelling and prediction of hydrological processes, and in the monitoring of the effect of catchment management policies. A primary concern of the study is the collection of sufficient and sufficiently accurate information to model hydrological processes. Remote sensing, in combination with conventional point source measurement, can be a valuable source of information, but is often overlooked by hydrologists, due to the cost of acquisition and processing. This thesis describes a number of cost effective methods of acquiring remotely sensed imagery, from airborne video survey to real time ingestion of meteorological satellite data. Inexpensive micro-computer systems and peripherals are used throughout to process and manipulate the data. Spatial information systems provide a means of integrating these data with topographic and thematic cartographic data, and historical records. For the system to have any real potential the data must be stored in a readily accessible format and be easily manipulated within the database. The design of efficient man-machine interfaces and the use of software enginering methodologies are therefore included in this thesis as a major part of the design of the system. The use of low cost technologies, from micro-computers to video cameras, enables the introduction of water resources information management systems into developing countries where the potential benefits are greatest.