Road extraction from high-resolution remote sensing image is an important and difficult task. The road-extraction method, which uses the integration shape features and the improved Hough transform, is proposed in this paper. Firstly, the image is segmented, and then the linear and curve roads are obtained by using several object shape features. Secondly, the step of road extraction is using the improved Hough transform method to deal with the road targets. Finally, the extracted roads are regulated by combining the edge information. In experiments, the images including the better gray uniform of road and the worse illuminated of road surface were chosen, and the results prove that the method of this study is promising.

No abstract

This paper aims to detect the defect of fabric under different illumination in texture industry. Although the definitions of fabric images are different under different illumination, they all have strong cyclicity and directionality. So a detection method is presented based on Fourier-polar coordinate processing in this paper. In frequency domain, the spectrum energy is transformed to the polar coordinate. Then in the polar coordinate, the sum of energy is added up at the same angle, and the main direction of image texture is found according to the energy sum. Salience of defect is enhanced by lowering the energy on the main direction. Experiments show that the method for detecting fabric defects with high accuracy can meet the requirements of real-time detection under different illumination.