Weiwei Jiang

Bio: Weiwei Jiang is an academic researcher from Shanghai University. The author has contributed to research in topics: Canny edge detector & Deriche edge detector. The author has an hindex of 1, co-authored 1 publications receiving 4 citations.

An Adaptive Edge Detection Algorithm Based on Improved Canny

Abstract: Edge detection is the key to image processing and has a significant impact on the high level of description, classification and matching of subsequent images. The traditional Canny algorithm requires human intervention in the selection of Gaussian function and its fixed parameters. To solve these problems, an improved algorithm based on Canny algorithm is proposed in this paper. The approach introduces the edge preserving filter to replace the original Gaussian filter, and calculates the magnitude and direction of image gradient with a new designed templates from x direction, y direction, and two oblique directions (45°, 135°). Meanwhile, the Otsu algorithm is used to calculate the thresholds, which avoids the problem that the thresholds need to be set repeatedly. The proposed method is successfully applied to the metal plate detection system. Experimental results show that the algorithm has good performance in bright and dark domains.

Research on image text recognition based on canny edge detection algorithm and k-means algorithm

Abstract: The latest research in the field of recognition of image characters has led to various developments in the modern technological works for the improvement of recognition rate and precision. This technology is significant in the field of character recognition, business card recognition, document recognition, vehicle license plate recognition etc. for smart city planning, thus its effectiveness should be improved. In order to improve the accuracy of image text recognition effectively, this article uses canny algorithm to process edge detection of text, and k-means algorithm for cluster pixel recognition. This unique combination combined with maximally stable extremal region and optimization of stroke width for image text yields better results in terms of recognition rate, recall, precision, F-score and accuracy. The results show that the correct recognition rate is 88.3% and 72.4% respectively with an accuracy value of 90.5% for the proposed method. This algorithm has high image text recognition rate, can recognize images taken in complex environment, and has good noise removal function. It is significantly an optimal algorithm for image text recognition.

Non-motor vehicle illegal behavior discrimination and license plate detection based on real-time video

Abstract: In recent years, automatic detection technology of motor vehicles has developed rapidly. However, due to some interference problems, there is little research on the automatic detection technology of non-motor vehicle. This paper proposes a method which based on the experience and results of actual projects for discriminating non-motor vehicles in real-time video, detecting and recognizing license plates. The algorithm and steps will be described in detail. The image difference method and fractional differential method are combined to extract the contours of dynamic non-motor vehicles in the video, and the methods are used to filter the contours of non-motor vehicles using skin color filter detection and geometric feature discrimination to assist the training of cascade neural networks. Non-motor vehicle license plate image clustering is detected in the extracted contour by clustering the background color of the license plate and determining the ratio of the rectangle surrounding the license plate. The boundary expansion method is used in combination with the Faster R-CNN (Faster-Convolutional Neural Networks) to train the model, and then BP (Back propagation) neural network is used to identify characters in the target area. And a trajectory tracking method is proposed to automatically determine whether non-motorized vehicles have violations.

Lane Line Edge Detection Based on Improved Adaptive Canny Algorithm

Abstract: With the development of autonomous driving technology, effective access to information on lane lines is of great significance for the decision-making of unmanned vehicles. The Canny operator is widely used in identifying the edge of the lane line, but the adaptive ability and anti-interference ability of the traditional Canny operator are flawed. In this paper, the lane line edge detection algorithm of the adaptive Canny algorithm is improved, and the adaptive median filtering and morphological closing operation are used to prevent the edge information from weakening while using multiple directions to calculate gradient magnitude. Finally, iterative method is used to improve the OSTU adaptive algorithm to determine its high and low thresholds. The test results show that the improved algorithm can effectively improve the accuracy, speed and anti-interference ability of the detection.

Implementation and Optimization of Template Matching Algorithm for FT-M6678

Abstract: Feiteng FT-M6678 (hereinafter referred to as M6678) DSP is a multi-core highperformance DSP with completely independent intellectual property rights. M6678 adopts the Harvard architecture and the new KeyStone multi-core architecture that store instructions and data separately. In general, the multiple classic image processing classic algorithms including correlation matching are not efficient for M6678 DSP architecture. To promote the application of domestic DSP chips in the field of image processing and artificial intelligence. Here, the correlation matching algorithm based on the correlation coefficient is transplanted to the M6678 platform, combining the algorithm characteristics and the architecture characteristics of the target platform, performance optimization in terms of parallelism and locality is carried out. The performance of the program is obviously improved, and the unique computing resources of the platform are more fully utilized, which is meaningful for the transplantation and optimization of other image processing algorithms on the platform.

Analysis of Coconut Tree trunk for the Identification of Cracks and Holes Using Image Processing

Abstract: Coconut Harvesting is one of the marketing crops in India. Some of the factors affecting coconut yielding are pest attacks such as red palm weevil, Stem bleeding through developed cracks. The aging of the coconut trees builds cracks, holes on the coconut trucks. Pest attacks, Cracks, and holes development affect the growth of the coconuts and results in reduced cultivation of coconuts. By identifying and controlling pest attacks, we can achieve better crop production. This paper aims to identify and spot the cracks and holes on coconut trunks using image processing techniques. In our proposed model, Cracks and holes identification uses different image preprocessing steps and four edge detection methods, and the crack is spotted using the contouring method on morphological Images. The algorithm proposed is tested on 50 real-time coconut trunk images collected from the Guntur, Prakasam districts of Andhra Pradesh, and Kollam district of Kerala in India. In this paper, we compare four edge detection models with multiple preprocessing steps required to is identify Cracks and Holes. Canny provides 96%, Prewitt with 89%, Sobel operator with 70%, Laplacian with 65% of accuracies