No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

One of the fundamental problems in the field of computer vision is the task of classifying objects, which are present in an image or sequence of images, based on their appearance. This task is commonly referred to as the object recognition problem. A system designed to perform this task must be able to learn visual cues such as shape, colour and texture from examples of objects presented to it. These cues are then later used to identify examples of the known objects in previously unseen scenes. The work presented in this thesis is based on a statistical representation of shape known as a pairwise geometric histogram which has been demonstrated by other researchers in 2-dimensional object recognition tasks. An analysis of the performance of recognition based on this representation has been conducted and a number of contributions to the original recognition algorithm have been made. An important property of an object recognition system is its scalability. This is the. ability of the system to continue performing as the number of known objects is increased. The analysis of the recognition algorithm presented here considers this issue by relating the classification error to the number of stored model objects. An estimate is also made of the number of objects which can be represented uniquely using geometric histograms. One of the main criticisms of the original recognition algorithm based on geometric histograms was the inability to recognise objects at different scales. An algorithm is presented here that is able to recognise objects over a range of scale using the geometric histogram representation. Finally, a novel pairwise geometric histogram representation for arbitrary surfaces has been proposed. This inherits many of the advantages of the 2-dimensional shape descriptor but enables recognition of 3-dimensional object from arbitrary viewpoints.

This thesis addresses the problem of automatic detection and identification of blemishes in digital images of potatoes. Potatoes are an important food crop, with clear unblemished skin being the main factor affecting consumer preference. Potatoes with defects, diseases and blemishes caused by otherwise benign (to human) skin infections, are strongly avoided by consumers. Most potatoes are sorted into dfferent grades by hand, with inevitable mistakes and losses. The currently deployed computer vision systems for sorting potatoes require manual training and have limited accuracy and high unit costs. A further limitation of typical machine vision systems is that the set of image features for pattern recognition has to be designed by the system engineer to work with a specific configuration of produce, imaging system and operating conditions. Such systems typically do not generalise well to other configurations, where the required image features may well differ from those used to design the original system. The objective of the research presented in this thesis is to introduce an automatic method for detecting and identifying blemishes in digital images of potatoes, where the presented solution involves classifying individual pixels. A human expert is required to mark up areas of blemishes and non-blemishes in a set of training images. For blemish detection, each pixel is classified as either blemish or non-blemish. For blemish identification, each pixel is classified according to a number of pre-determined blemish categories. After training, the system should be able to classify individual pixels in new images of previously unseen potatoes with high accuracy. After segmenting the potato from the image background, a very large set of candidate features, based on statistical information relating to the colour and texture of the region surrounding a given pixel, is first extracted. The features include statistical summaries of the whole potato and local regions centred on each pixel as well as the data of the pixel itself. Then an adaptive boosting algorithm (AdaBoost) is used to automatically select the best features for discriminating between blemishes and non-blemishes. The AdaBoost algorithm (Freund and Schapire, 1999) is used to build a classifier, which combines results from so-called "weak" classifiers, each constructed using one of the candidate features, into one "strong" classifier that performs better than any of the weak classifiers alone. With this approach, different features can be selected for different potato varieties, while also handling the natural variation in fresh produce due to different seasons, lighting conditions, etc. For blemish detection, the classifier was trained using a subset of pixels which had been marked as blemish or non-blemish. Tests were done with the full set of features, "lesion experiments" were carried out to explore the impact of removing specific feature types, and experiments were also carried out on methods of speeding up classification both by restricting the number of weak classifiers and restricting the numbers of unique candidate features which can be used to produce weak classifiers. The results were highly accurate with visible examples of disagreement between classifier output and markup being caused by human inaccuracies in the markup rather than classifier inaccuracy. For blemish identification, a set of classifiers were trained on subsets of pixels marked as each blemish class against a subset of pixels drawn from all other classes. For classification, each pixel was tested with all classifiers and assigned to the classifier which returned the highest confidence of a positive result. Experiments were again performed with methods of speeding up classification as well as lesion experiments. Finally, to demonstrate how the system would work in an industrial context, the classification results were summarised for each potato, providing a high overall accuracy in detecting the presence or absence of significant blemish coverage for each blemish type.

No description available.

In both RMCS calibration and ACS calibration, the corresponding efficiency and robustness are tested by simulation and real experiments. In the real experiment of ACS calibration, the intrinsic and extrinsic parameters of the ACS are obtained simultaneously by our calibration procedure using the same image sequences, no extra data capturing step is required. The corresponding trajectory is recovered and illustrated using the calibration results of the ACS. Since the estimated translations of different cameras in an MCS may scaled by different scale factors, scale factor estimation algorithms are proposed for non-overlapping view RMCS calibration and ACS calibration respectively. To our knowledge, we are the first to study the calibration of ACS. / In this thesis, we focus on developing robust methods for the MCS calibration problems. In particular, we make two contributions. Firstly, we developed a novel extrinsic calibration method for the non-overlapping view Rigid Multiple Camera System (RMCS) using the kinematic information of the RMCS. The input are only the images captured when the non-overlapping RMCS is moved in an environment with enough static feature points. This assumption is true in many vision tasks such as SFM (Structure from Motion), SLAM (Simultaneous Localization and Map). The output is the extrinsic parameters of the cameras of the RMCS. / Multiple Camera Systems (MCS) have been widely applied in many vision applications and attracted much attention recently. Both intrinsic and extrinsic parameters of an MCS are needed to be calibrated before it is used. / Secondly, we proposed to solve the calibration of a particular model of non-rigid Multiple Camera System, namely, Articulated Camera System (ACS). In an ACS, the cameras are fixed on articulated arms with joints, the relative pose between them may change. Two ACS calibration methods are proposed. In the first approach, we assume the cameras have overlapping views. It uses the feature correspondences between the cameras in the ACS. In the second approach, we assume the cameras have no overlapping view. It requires only the ego-motion information of the cameras and can be used for the calibration of the non-overlapping view ACS. In both methods, the ACS is assumed to have performed general transformations in a static environment. / Chen, Junzhou. / Adviser: Kin Hong Wong. / Source: Dissertation Abstracts International, Volume: 70-06, Section: B, page: 3594. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2008. / Includes bibliographical references (leaves 102-110). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese. / School code: 1307.