Papers by Keyword: Edge Detection

Abstract: A novel 3-D edge detection methodology is developed to resolve the edge ambiguity problem encountered in 3-D optical surface profilometry employing digital image correlation (DIC). DIC has been surged as a full-field measurement technique for in-plane and out-of-plane dynamic mechanical structure analyses. However, up to date, one of the key issues in DIC is still remained in boundary edge detection since a surface edge is not detectable between two discrete neighboring height jumps due to optical diffraction. Generally, it is common to observe undesirable noisy measured data along surface edges in traditional DIC-based surface 3-D profilometry. To resolve this, a novel random speckle images processing method is established by proposing a new algorithm by employing the multiple symmetric partial template model to determine best edge location with accurate height reconstruction. A theoretical simulation on a pre-calibrated circle target was performed to verify the feasibility of the methodology. Some experiments on real industrial objects having various surface reflective characteristics were implemented to verify its capability on accurate detection of industrial objects having discrete surface edges. From its preliminary evaluation on measurement accuracy, it is found that the maximum measured error on critical dimension can be controlled within less than 6.0% of the overall measuring range while one standard deviation can be kept within less than 1.2%.

Abstract: As is well known, the phase-shifting interferometry techniques allow to reach longitudinal resolution to ~ 0.1 nm, but the value of lateral resolution remains at the level of ~ 1 mm. For providing of high lateral resolution of linear measurements in the interference microscope profilometer it was proposed to use a position detection sensor of sharp edge. Principle of sensor’s measurement is based on registration of laser spot intensity scattered by the measurement sample surface under displacement of sample in the lateral direction. The paper shows the prototype scheme of measurement system containing the Linnik interferometer used for surface nanorelief measurement and a position detection module of sharp edge. Measurement process and experimental results are presented. The combining of measurement results performed by the Linnik interferometer and a position detection sensor of sharp edge can allow us to precisely (better then diffraction limit) define the position of sharp edge on the reconstructed surface nanorelief.

Abstract: The traditional first-order differential operator is under the influence of the Gaussian noise, therefore, it often implement boundary extraction after average filtering. But the filtering process would often smooth the details of some directions of image too much, so that the edge cannot be extracted correctly. To solve this problem, this paper puts forward the edge detection algorithm based on edges keep, to determine the keeping direction of the edge through matching different directions’ edge template. Instead of average filtering process, it can improve the performance of traditional operator, and provide the simulation results. Experimental results show that the algorithm can eliminate noise, and at the same time, keep more edge information of the image.

Abstract: Lane detection system for the driver of the car is an important issue for the inquiry as a platform for safe driving experience. Implementation of this system is trying to investigate the possibility of traffic accidents, monitor the efficiency of the movement and position of the car contributes to the development of autonomous navigation technology. The purpose of this study is to get the best selection of banks in a better Hough transform technique to detect lane roads using edge detection techniques. For this study, Canny, Sobel and Prewitt edge detection is used as a trial. Selection of the best edge detection was using neural network techniques. Improved Hough Transform is used to extract features of a structured road. Point area near the straight line model adopted to accelerate the speed of calculation data and find the appropriate line. Prior knowledge is used in the process of finding a path to efficiently reduce the Hough space efficiently, thereby increasing the resistance by increasing the processing speed. Experiments provide good results in detecting straight and smooth fair curvature lane on highway even the hallways are painted shadows. Data from the lane highways have been taken in video format. Experiments have been done using an edge detection technique of choice in each scenario, and found that the best method of producing high accuracy of detection is to use intelligent edge detector. In this way, other people will be the best in certain cases scenarios lane highway.

Abstract: The Image processing technique incorporates human perception and intelligence which makes this field more interesting to the research community. The edge detection process is the most important step in image recognition system. In this paper a simple three dimensional model is created by taking the best edge detected image followed by comparison with various edge detection techniques using Labview software. HereCatia model of spur gear isdrawn by observing and analysing best suited edge detected image in order to make the design more precise in edges and geometry and also to make object recognition simple and further scope is given to design a Catia model using dimensional parameters with the help of vision assistant tool.

Abstract: Generally after a cylindrical steel bar is made from hot forming process, its contour will look like ‘U’ or ‘wavy’ topography. The preformed steel bar after steelmaking process needs to remove the outside oxide layer by lathe or peeling machine (commonly known as stripping). As a consequence, the deformation of steel bar will cause subsequent problems, such as eccentric rotating of steel bar as the lathe is running. Although it is necessary to remove the oxide layer, the deformation will increase waste materials. To prevent the problems, a straightening machine is usually used to straighten the cylindrical steel bar before the pickling process. However, most current straightening machines cannot measure cylindrical steel bar contour on-line instantly. This study proposed a test method for measure the contour of a large cylindrical steel bar based on multi-line structured light and machine vision. Break line method were performed to measure the contour of the cylindrical steel bar. The experimental result shows that the measurement error is in the range of ±0.4 mm after calibration. The study result concludes that the proposed method can be applied to straightening machine for real-time online measurement to improve straightening efficiency.

Abstract: This paper is concerned with an algorithm of determining a position of an object and its borders in an X-Ray image. The algorithm is based on a preliminary estimation of a histogram of given image. The information retrieved from estimation provides complementary parameters for further edge detection. As a result, this approach allows reducing the processing time. The future research and development of the algorithm will be aimed at object tracking in real-time systems.

Abstract: Road image edge detections are sensitive response to light intensity, the whole mean gray
value of road image changed significantly under different illumination. Using one algorithm can not
be effectively detected to road edge line. From the analysis of characteristic situation of different
light road image, we put forward a algorithm of road edge line detection under different illumination.
The example studied showed that this algorithm could obtain more ideal image edge from the road
gray image, and certain anti-noise performance could be generated.

Abstract: Quality of steel rail affects directly the running state of locomotive. According to detection problem of rail head’s abrasion loss, a feasibility study was done that collecting image through CCD camera. Computing models of abrasion detection were established by MATLAB soft. The key question about image location was studied using the method of edge detection and image filling. Results indicate that the image detection system could obtain the abrasion loss accurately in non-contact condition. This method could provide a reference for the development of loss prevention and new detection technology.

Abstract: Based on the principle of machine vision technology, we designed a methodto detect the outline dimensions of automotive airbag quickly and accurately. We Used CCD camera obtain the airbag image, through the image processing method ofsmooth filtering andgray-scale transformationto complete pre-processing, finally applied Canny edge detection operator to extract the boundary of the airbag contour features,and then took the template matching methodto detect assemble error of the airbag image whether meet the requirement.The results show that the detection method have a higher precision, and the time is very short, it can improve the sampled positioningerror detection for the all checks image recognition detection, suitable for application in real-time online detection of airbag assembly line.