Kurt Kubik

Using the Dempster–Shafer method for the fusion of LIDAR data and multi-spectral images for building detection

Abstract A method for the detection of buildings in densely built-up urban areas by the fusion of first and last pulse laser scanner data and multi-spectral images is presented. The method attempts to achieve a classification of land cover into the classes “building”, “tree”, “grassland”, and “bare soil”, the latter three being considered relevant for the subsequent generation of a high-quality digital terrain model (DTM). Building detection is accomplished by first applying a hierarchical rule-based technique for coarse DTM generation based on morphological filtering. After that, data fusion based on the theory of Dempster–Shafer is used at two different stages of the classification process. We describe the algorithms involved, giving examples for a test site in Fairfield (New South Wales).

Automatic building detection using the Dempster-Shafer algorithm

An approach and strategy for automatic detection of buildings from aerial images using combined image analysis and interpretation techniques is described in this paper. It is undertaken in several steps. A dense DSM is obtained by stereo image matching and then the results of multi-band classification, the DSM, and Normalized Difference Vegetation Index (NDVI) are used to reveal preliminary building interest areas. From these areas, a shape modeling algorithm has been used to precisely delineate their boundaries. The Dempster-Shafer data fusion technique is then applied to detect buildings from the combination of three data sources by a statistically-based classification. A number of test areas, which include buildings of different sizes, shape, and roof color have been investigated. The tests are encouraging and demonstrate that all processes in this system are important for effective building detection.

Suitability Analysis of Techniques for Flaw Detection in Textiles using Texture Analysis

Abstract: This paper addresses issues conerning image analysis algorithms for the visual quality inspection of textile fabrics. An overview of a number of flaw detection techiques and analysis of their suitability for detecting the presence of weaving defects is presented. The flaw detection algorithms are based on a local analysis of spatial and spatial-frequency domain features, exploring both the statisical and structural texture analysis of the defective image samples. The algorithms are compared using two criteris: accuracy and computational efficiency. Four methods are applied to the flaw detection problem, and their accuracy is assessed experimentally on a number of common textile defects in this preliminary comparison. The methods are: spatial grey level co-occurrence; normalised cross-correlation; texture blob detection; and spectral approeaches. Of these mehods, the correlation approch appears to be the most promising candidate for a real-time, high accuracy flaw detection algorithm.

Passive bistatic radar sensing with LEOS based transmitters

In this paper we describe the concept of a bistatic SAR system which involves a purpose built stationary ground based receiver but which utilizes existing or proposed low earth orbit satellite (LEOS) communication systems, such as Globalstar or ICO as the transmitting signal sources. In general, such a system has a 3-D configuration (in contrast to the 2-D bistatic SAR systems examined previously in the literature). We analyse this system. With earth observation applications in mind, we characterize quantitatively the 2-D resolving capability within the ground plane. This is demonstrated via a meaningful simulation.

Texture characterization of compressed aerial images using DCT coefficients

We test the performance of a texture feature constructed from the variance of the first eight AC Discrete Cosine Transform (DCT) coefficients of JPEG compressed images. We break the image into sub-images, consisting of many 8*8 blocks, and them calculate the variance of each DCT coefficient across the sub-image. We evaluate the texture feature at two different image resolutions, and at three different quality factors. In our high resolution image a pixel covered a square of side 4 cm on the ground. Our low resolution image was generated by subsampling. Representative feature vectors were generated for give subjectively identified textures, by averaging a small training set. Each sub-image was then classified according to the representative feature vector closest in feature space. Compression ratio had little effect on the classification result in our study. However image resolution significantly altered the classification result. Classification correlated much more closely to a subjective classification for the low resolution image. Feature vectors also fell into much more clearly defined clusters at the lower resolution. Although more research is required across different photo-scales and sets of images, we conclude that texture features generated from compressed JPEG images have potential for content-based image retrieval based on texture.

RELATIVE AND ABSOLUTE ORIENTATION BASED ON LINEAR FEATURES

Abstract In industrial metrology, when measuring the actual shapes of industrial tools and jigs, it is often impossible to define unique measuring points on an object. Straight edges (or circular areas) are however, commonly available. In this paper we described the mensuration procedure using these edges instead of well defined object points. We described the method of computation and the mensuration procedure, with the minimum number of lines necessary to calculate the problem successfully.

Optimal textural features for flaw detection in textile materials

This paper examines the problem of quality control and defect identification in woven textile fabrics by introducing an improved method for texture description. The approach is based on spatial gray level dependence methodology and addresses the issue of optimal parameter selection for deriving the maximum textural information. We introduce the use of the /spl chi//sup 2/ significance test on elemental feature matrices in order to obtain higher per feature texture description and hence improve the capture of defects in the underlying textile pattern.

Defect detection in textile materials based on aspects of the HVS

The problem we address in this paper is that of detecting flaws in woven textile fabrics. This task is currently performed by human inspectors with a maximum detection rate of only about 80%. An Automated Visual Inspection System (AVIS) is potentially a more reliable, objective, time and cost effective solution. Using computer vision it detects irregularities in homogeneously structured images of textile fabrics. However, its successful performance is largely dependent of the choice of an accurate and robust texture analysis algorithm. The technique of blob detection described in this paper is representative of a structural texture analysis approach and it accounts for aspects of the human visual system (HVS) in detecting large varieties of textile flaws. Measures of texture discrimination based on psychophysical experiments are used to indicate the levels of perceivable differences in blob attributes indicating the presence of defects. Defects are detected by observing a threshold of acceptable differences in the properties of blobs based on human perception.

Power budget study for passive target detection and imaging using secondary applications of GPS signals in bistatic radar systems

Signals from satellite systems like GPS, IRIDUM and Globalstar that are reflected from moving or stationary objects are utilised for their detection in a synthetic aperture bistatic radar system (SAR). The movements of the satellite and its position at different coordinates with respect to time can serve as a base for a synthetic aperture. This paper focuses on the development of a system based on the analysis of indirect signals, in particular GPS signals. The main concern in this study is the low power of the GPS signal at the earths surface (typically about -160 dB). Due to the nature of the reflector/target, some power will be absorbed, and hence result in a further reduction in signal strength reflected by the target. However, the other equally important concern here is the ground and sea clutter, which is 18-22 dB stronger than the direct signal level. Here we present theoretical results of both the maximum detection range and land clutter contributions of the air target detection by bistatic SAR that utilizes the existing GPS satellite as the transmitting signal source.

Analytical solution for target location using bistatic multi-transmitter and multi-receiver techniques

The 3D bistatic radar solution determines the coordinates of a target from bistatic radar measurements, where the transmitter and receiver are at different positions. This involves the solution of non-linear equations. The conventional approach is linearising these equations. In this paper, we propose a direct solution of the non-linear equations. This allows the evaluation of the multiple solutions, study of singularities and the design of geometric configurations which lead to unique solutions. The solution is demonstrated for satellite based transmitters which are used as sources of opportunity.