A Cross-Platform OpenVX Library for FPGA Accelerators

In computer vision, edge detection is a hot research area in which Canny operator is a typical algorithm. Canny operator has preferable anti-noise ability. However the edge based on Canny operator is not consecutive. GVF Snake model is used widely in image segmentation. But there are problems in convergence processing to boundaries of some medical image because of noise. This paper presents a new segmentation algorithm to medical image. First, rough edge is got by Canny operator, and then thinning method based on mathematical morphology is adopted to get edge map as foundation of GVF Snake model. This method solves the problem that the edge based on Canny operator is not consecutive. And it improves GVF Snake model's anti-noise ability. Experiments indicate that the new algorithm can improve snake model's ability to segment the complicated image.

As an indispensable measuring instrument, the pointer instrument has been widely used in the fields of industrial production. While the manual instrument identification has problems of higher labor intensity, higher rate of misjudgment and lower production, we take advantage of computer vision to deal with the images of pointer instrument dials. Two algorithms are proposed for image edge detection: structured random forests edge detection and canny edge detection, Hough transform is used to recognize the straight lines of the pointer, then the automation of instrument identification can be realized. Our results show that the accuracy of the dials printed by the two methods can meet the production requirements.

Feature detection is a fundamental technique in broad fields of image processing, pattern recognition and computer vision. A digital image in general contains objects, edges, noises and background. Critical changes in properties of objects can be captured via detecting sharp variations in image brightness. The edges can be detected via numerous approaches on a basis of image intensity changes. Edge broken and false detection are typical problems using classical methods, which will result in information loss and feature deformity. The notion of optimization is thus introduced into edge detection. The Canny edge detector and Ant Colony Optimization (ACO) detector are among the most successful and effective approaches for edge detection. The Canny edge detector is designed to capture edges by searching local optima of the gradient of the intensity. It is susceptible to noises presenting on the raw images, so details of images could be slightly changed when Gaussian smoothing is applied. To improve accuracy, the adaptive edge tracing scheme is proposed. On the other hand, artificial intelligence has also been introduced. Being one of metaheuristic optimization approaches, the evolutionary computing oriented ACO becomes a promising approach for feature capturing without necessity of smoothing filters. Selection of maximum intensity difference as the path visibility function for ACO will contribute better to generate true edges and avoid false edges. Both the adaptive Canny edge detection and enhanced ACO are proposed in this article. Comparative studies are also conducted to evaluate the edge detection qualities. The outcomes are analyzed and evaluated from both qualitative and quantitative points of view, where merits and drawbacks of the two schemes have been indicated.

Human vision can do amazing things such as recognizing people or objects, navigating through obstacles, recognizing the mood in a scene, and imagining stories. To do mimicry of the human vision, the computer requires a sensor that functions like the human eye and a computer program that serves as a data processor from the sensor. Computer vision is the science that uses image processing to make decisions based on images obtained from sensors. In other words, computer vision aims to build an intelligent machine that can see. Computer vision can be used to detect skin diseases, for example, to detect disease Shingles (Herpes Zoster), Hives (Urticaria), Psoriasis, Eczema, Rosacea, Cold Sores (Fever Blisters), Rash, Razor Bumps, Skin Tags, Acne, Athlete's Foot, moles, Age or Liver Spots, Pityriasis Rosea, Melasma (Pregnancy Mask), Warts, and Seborrheic keratoses. Prewitt, Sobel, Roberts, and Canny operator are used to detect the edges of one or more objects. Then the results will be match with the results of edge detection image data base to determine the type of disease using Scale invariant Feature Transform (SIFT) algorithm. Skin Disease Detection Expert System will be implemented with C++ programming language, IDE MS Visual Studio 2010 and OpenCV 2.4 library. Keywords— computer vision, edge detection, SIFT algorithm, skin disease

The article discusses the importance of selfdriving cars to improve road safety and reduce the number of accidents caused by human error. Self-driving cars not only reduce human error but also help reduce driver fatigue. We further explore the use of computer vision in autonomous cars, with previous research relying on deep learning algorithms with LiDAR sensors which can be expensive. The authors propose a more cost-effective approach using simple computer vision algorithms such as color space transformation, Canny edge detection, and Hough line transformation to detect lane lines and steer the car accordingly. This approach requires less operational hardware and can be implemented using affordable boards like Raspberry Pi and Nvidia Jetson Nano. The article also highlights the reconstruction of a remote-controlled car that had a 95% accuracy using a certain set of parameters was a tool for understanding autonomous cars better.

A comprehensive study on lane detecting autonomous car using computer vision

This investigation delves into an advanced method for enhancing robot autonomy by integrating hand-drawn sketches with computer vision systems to optimize route planning effectively. The procedure starts with the capture of image data, which is then processed through multiple stages until the sketches are turned into exact coordinates that guide a robot. Canny edge detection, a specific computer vision technique, is employed to detect the edges of the drawn lines, facilitating precise and dependable navigation for autonomous robot operations. Initially, the image is subjected to noise reduction and edge enhancement before converting the sketch lines into Cartesian coordinates. This step is succeeded by point filtering and ordering to ascertain accurate path adherence by the robot, with no coordinates overlooked. Scale adjustments are also made to match digital image coordinates with the real-world setting, thereby improving the robot’s interaction based on the received visual inputs. The program has been tested in Robotic Operating System 2 (ROS) using MoveIt2 and also in RoboDK software, where it was used to verify the results. It is important to note that path planning, which focuses on determining an optimal route from start to finish, is a subset of motion planning, which also considers the dynamics and kinematics of the robot's movement along that route.

Harris corner detection algorithm has been widely used in many computer vision allocations. However, it has low efficiency and accuracy, poor noise immunity and needs to set an artificial threshold. In this paper, an improved algorithm based on Canny edge detection and gray difference preprocessing is proposed. Firstly, Canny edge detection and gray difference preprocessing are used for corner prescreening to improve the detection efficiency, anti-noise, and rotation invariance. Secondly, non-maximum suppression is applied to the screened corners to reduce the number of false corners. Finally, the average of adjacent points method is used to solve the problem of corner cluster, and the detection results are compared the measurement accuracy is improved to sub-pixel level. Experimental results indicate that the proposed algorithm can accurately extract the corners in the image and remove the false corners and corner clusters. It achieves superior performance than existing methods.

This paper presents the design and evaluation of Meghadoot - a hybrid wireless network architecture that provides improved services, such as better connectivity among the users, efficient routing, and access to the Internet. Meghadoot architecture combines the advantages of both single-hop and multi-hop wireless networks. Meghadoot is extremely useful (a) as an alternative to existing wired or wireless last mile solutions, (b) to establish a network in a residential complex or in a university campus. The Meghadoot architecture has been implemented and tested in both the NS2 simulator and in the Linux kernel. Both, experimental and simulation results show an improved performance for Meghadoot, in terms of reduction in control overhead and increase in throughput over that of the existing approaches.

Smoothed differentiation filters for images

Egg crack is one the main challenges that should be identified before sending it to market. In one hand, cracked eggs have strong potential in taking bacterial contamination and also they cause negative effects on the other intact eggs at packages so that this contamination endanger human’s health. The objective of the current study was to propose a computerized method as an accurate, non- destructive and a fast method to identify the eggshell cracks. In order to detect the defects, the Hough transformation as a confident and qualified method with having the advantage of description based feature was used in determining the line in the images with assuming cracks as lines. The dataset consisted of 80 eggs which were included 45 healthy and 35 cracked eggs where taken images under controlled conditions. The cracks on the egg shells were identified by applying common preprocessing operations, a Canny edge detector and finally Hough transform. In the analysis section, the linear discriminant analysis was used to classify healthy samples from cracked ones. The results demonstrated satisfactory of the proposed approach in identification and classification of intact and cracked eggs so that we were able to reach 90.1% of accuracy in correct identification. The time for identifying the cracks in each egg was obtained 0.7 seconds.

In recent years, the demands for LED are increasing. For testing the quality of LEDs, a lot of LED probes are necessary, so the high precision and efficient methods are paid more attention by industrial applications. This paper is focused on the measurement of the angle and the radius of a LED probe by computer vision. In previous paper, we proposed an effective method based on Canny edge detection and curve fitting method with iteration. In this paper we add a new sub-pixel edge detection method: partial area effect. We improve the preciseness of angle from error 2.3% to 1.43% and enhance the accuracy of radius more than 30%.

Autonomous vehicles necessitate the integration of advanced technologies such as computer vision and deep learning to comprehend and navigate their surroundings. A crucial yet challenging component of this integration is the accurate detection of lanes, which can be influenced by a multitude of varying lane characteristics and conditions. This research undertakes a comparative analysis of lane detection methodologies, explicitly focusing on traditional image processing techniques and Convolutional Neural Networks (CNNs). The evaluation utilized a sample of 500 images from the CULane dataset, which encompasses a diverse range of traffic scenarios. Initially, a method incorporating Gaussian blurring, Canny edge detection, and Hough line transformation was examined. Despite its efficiency, operating at 30 frames per second, this approach exhibited a high error rate (average Mean Squared Error (MSE) of 0.537), which is attributable to the loss of critical image details during the preprocessing stage. Subsequently, the performance of a fine-tuned YOLOv8 model, trained on a reformatted version of the CULane dataset was assessed. The combination of object detection and subsequent Hough transformation yielded high accuracy, demonstrating the model’s ability to learn and identify relevant lane features. The deep CNNs demonstrated superior performance over classical image processing techniques in terms of lane detection accuracy, thereby underscoring their potential applicability within the realm of autonomous vehicle technology

Quantitative study of starch swelling capacity during gelatinization with an efficient automatic segmentation methodology

On the Canny edge detector