摘要：The classical edge detectors work fine with the high quality pictures, but often are not good enough for noisy images because they cannot distinguish edges of different significance. The paper presented a novel approach to multiscale edge detection for noisy images using wavelet transforms based on Lipschitz regularity coefficients and a cascade algorithm. The relationship between wavelet transform and Lipschitz regularity was established. The proposed wavelet based edge detection algorithm combined the coefficients of wavelet transforms along with a cascade algorithm which significantly improves the result. The comparison between the proposed method and the classical edge detectors was carried out. The algorithm was applied to various images and its performance was discussed. The results of edge detection of contaminated images using the proposed algorithm show that it works better than the classical edge detectors.

摘要：in many industrial situations, edge detection of 3-D objects is frequently *** examples are found in dimension control and machine vision applications .As technology advances, the capability of computers has greatly improved. This paper describes our system used to inspect objects and the method used to find the edges for those objects with planar and cylindrical faces from a huge set of cloudy points. First the systems are based on opto-electronic techniques and hence are non-contact. The system is designed for a small objects less than 2 x 2 x 2 cm. The edge resolution is based on the principle of spacial geometry of a triangle in the cartesian coordinate system. The system can be considered scanning linearly with a rotation stage of a mirror,on the other hand the object circumferentially rotate on a turntable. After obtained surface profile of 3-D objects,we then apply for a geometrical method to find the edges. The performance of the systems is demonstrated with different sizes test objects.

摘要：This paper describes a technology of dynamic display moving image by computer monitor,which is initially used in the design of tool detection system. The paper presents the hardware and software principie and edge detection process. The way of marking edge point coordinates and stability of moving image also is analyzed. The method reforms the conventional design of the 2-D vision detection system. Moreover,it facilitates the design of the systematic mechanical construction,is convenient to compile instrument systemsoftware,and realizes to detect and track display image simultaneously. By the work,the tool detection system is improved to practical application.

摘要：The serrated incisors of grasshopper [Chondracris rosea rosea （De Geer）] possess an advantageous capacity for cutting plant fiber. Inspired by this special geometrical structure of incisors, bionic saw blade was designed and manufactured. MATLAB software digital image processing technology was used to obtain outer margin profile from stereomicroscope pho- tograph of the serrated incisors. The outer margin profile of incisors was fitted and expressed by six-order polynomial function. To compare the cutting capacity of bionic and traditional saw blades, the internodes of dry corn stalks were cut perpendicularly. Cutting force-deformation characteristics were obtained by universal testing machine. The results of cutting experiments show that the maximum cutting force of bionic saw blade was 128.26 N, which is 15.87% lower than 152.45 N of traditional saw blade; the average cutting force of bionic saw blade was 51.56 N, which is 28.17% less than 71.78 N of traditional saw blade. Meanwhile, the cutting energy consumption of bionic saw blade was 8.95 J, which is 12.85% less than 10.27 J of traditional saw blade. Overall, the bionic saw blade can lead to noticeable reduction of the cutting force and energy. These results will be helpful for designing cutting elements of corn stalk harvesting, biomass size reduction and other processing machinery.

摘要：The Canny edge detector inevitably misses some important and obvious edges during contour extraction, which causes gaps in the contour. A geometric method to locate, measure and fill the gaps precisely is proposed. With the complete contour information, a convolution approach is presented, which utilizes an appropriate linear interpolation to resample the contour to calculate pointwise curvature. This approach distributes discrete points within a convolution window uniformly. It ensures a one-to-one correspondence between every point and its weight, thus the accuracy is guaranteed under this condition. A related parameter selection is also suggested. Experimental results show that the proposed methods are robust and accurate.

摘要：A method based on HSI color space is presented to solve the problem of circle detection from color images. In terms of the evaluation to the edge detection method based on intensity, the edge detection based on hue is chosen to process the color image, and the simplified calculation of hue transform is discussed. Then the algorithm of circle detection based on Canny edge detection is proposed. Due to the dispersive distribution of the detected result, Hough transformation and template smooth are used in circle detection, and the proposed method gives a quite good result.