China
Africa
Explore chapters and articles related to this topic
A fast detection method for tunnel surface defects based on video processing
Published in Daniele Peila, Giulia Viggiani, Tarcisio Celestino, Tunnels and Underground Cities: Engineering and Innovation Meet Archaeology, Architecture and Art, 2020
As there are gaps in each ring edge of collected images, the offline positioning can be achieved through edge detection. The classical edge detection method is to build an edge detection operator for a small neighborhood of pixels in the original images. Several commonly used edge detection operators include Prewitt, Roberts, Sobel, LOG, Laplace and Canny edge detection operator. Abdel Qader et al. (2003) compared crack detection performance of fast Haar transform, Fourier transform, Sobel filter and Canny filter in 25 defective concrete images and 25 healthy concrete images. And Fast Haar transform is the most accurate method, with a total accuracy rate of 86%,followed by Canny filter (76%), Sobel filter (68%) and Fourier transform (64%). However, the processing time is not considered in the analysis, and the error criteria in the binary image are not clear.
Basics of Image Processing
Published in Maheshkumar H. Kolekar, Intelligent Video Surveillance Systems, 2018
The Canny edge detector first smooths the image to reduce noise. Then it finds the image gradient to highlight regions with high spatial derivatives. After this, all those pixels in these regions that are not at maxima are supressed. Hysteresis is now used to further reduce the gradient array. Hysteresis is used to check the remaining pixels that have not been suppressed. Hysteresis uses two thresholds, and if the magnitude is below the low threshold (T1 $ T_{1} $ ), it makes the pixel as non-edge. If the magnitude is above the high threshold (T2 $ T_{2} $ ), it makes it edge. If the magnitude is between two thresholds, then it is set to zero unless there is a path from this pixel to a pixel which is an edge.
Image Segmentation
Published in Vipin Tyagi, Understanding Digital Image Processing, 2018
The Canny edge detector can detect edges in noisy images with good accuracy because it first removes the noise by Gaussian filter. It generates one pixel wide ridges as the output by enhancing the signal with respect to the noise ratio by the non-maxima suppression method. Overall, it gives better edge detection accuracy by applying double thresholding and edge tracking hysteresis method. But the Canny edge detector is time-consuming and it is difficult to implement the Canny detector and reach real time response speeds.
Determining the maximal inscribed rectangle of an irregularly shaped stone using machine vision
Published in International Journal of Computer Integrated Manufacturing, 2022
Yu-Ting Luo, Ching-Fang Chen, Syh-Shiuh Yeh
This study uses Canny edge detection to determine the edge of a stone image (region of interest shown in Figure 6). The steps involved in the Canny edge detection operation generally include Gaussian filtering for removing noise, computing the image intensity gradient magnitude and direction, and removing non-edge pixels (Canny 1986). Figure 8 shows the Canny edge detection result for the erosion image in Figure 6. The edge pixels of the stone image can be obtained using Canny edge detection; therefore, the edge pixels should be synthesized as a continuous edge contour using the contour determination method. The contour determination method employed in this study is topological structural analysis (Suzuki and Abe 1985), and Figure 9 shows the contour determination result for the edge detection image in Figure 8.
Coastline changes in mainland China from 2000 to 2015
Published in International Journal of Image and Data Fusion, 2022
Ying Zhang, Qinghua Qiao, Jia Liu, Huiyong Sang, Dazhi Yang, Liang Zhai, Ning Li, Xiaohui Yuan
The object-oriented classification method can extract the coastline by extracting the sea surface. However, for coastal farming areas, misjudgements are often made because of perennial water. In order to improve the accuracy of coastline-boundary extraction, we comprehensively utilised edge detection and object-oriented methods to extract the coastline. A canny edge detector is widely used because of its stability under noise interference and the reliability on detecting edges (Surhone et al. 2010). In this study, the canny operator was involved in the object-oriented segmentation link, which was used as one of the weights to improve the accuracy of sea–land segmentation, thereby improving the accuracy of coastline extraction. The specific technical route is shown in Figure 1.
Shadow detection for mobile robots: Features, evaluation, and datasets
Published in Spatial Cognition & Computation, 2018
Charles C. Newey, Owain D. Jones, Hannah M. Dee
First, to detect strong edges in the image, Canny edge detection (Canny, 1986) is applied to a grey-scale copy of the input image with a lower hysteresis threshold of 0.3 and an upper threshold of 0.6. These thresholds have been selected empirically for our image sets, as they were shown to filter weak edges caused by textured surfaces, while preserving the edges of shadow boundaries in the candidate images. Afterwards, a list of of edge normals is generated: To generate a list of normals to inspect, a contour-finding algorithm (Suzuki & Be, 1985) is applied to the edges image to create a list of contour points. Normals are selected at the center of two points. To ensure complex contours aren’t over-sampled, a minimum distance of 5 pixels between candidate normals is enforced. Pixel values alongare then sampled across each normal provide us withto get the profile of the edge to which that normal belongs.