37 results about "Contour set" patented technology

The application discloses a method and a device for recognizing text areas in an image. A specific embodiment of the method comprises: acquiring a color value and position information of each pixelin the image to be recognized; clustering the pixels based on the color value of each pixel, wherein the color values of the pixels in each pixel category are the same or similar; for each pixel category after clustering, determining the contour of each connected area constituted by the pixels in the pixel category to obtain a contour set; and merging the contours based on the color value and position information of each contour in the contour set to obtain respective text areas in the image. The embodiment improves the accuracy of text line sequence recognition in image character recognition.

The invention discloses a three-dimensional model search method based on mesh segmentation. The method comprises the following step of analyzing and constructing a segmentation field through a hierarchical spectrum, and particularly comprises the steps of judgment of concave vertexes, constructing of a Laplacian matrix, matrix decomposition, selection of low-frequency feature vectors, generation of sub feature vectors, weight calculation of the sub feature vectors, and constructing of an edge symbol matrix. Contour lines are sampled in the segmentation field and are grouped and merged through a grouping-merging algorithm to obtain a plurality of candidate contour sets, the final segmentation boundary is determined according to the weight of each contour line in the candidate contour sets, and three-dimensional models are automatically segmented. Three-dimensional model mixing feature description sub-matrixes are obtained by calculating the feature description sub-matrix of each segmentation block of the three-dimensional models, the similarity of the mixing feature description sub-matrixes of each three-dimensional model in a three-dimensional model database to be searched and a target three-dimensional model to be searched is calculated, the similarity values of the three-dimensional models are ranked and output from low to high, and the three-dimensional model searching is achieved.

The invention discloses a photovoltaic module glass crack method, device and apparatus and a storage medium. The module includes: binarizing an image of a photovoltaic module to obtain a binarized image; extracting a contour of the binarized image to obtain an initial contour set C; filtering all initial contours according to the area of the outer rectangle and the number of corner points to obtain the contour set G of the photovoltaic module; traversing other initial contours except the photovoltaic component contours in the initial contour set C, obtaining an initial contour contained in anyof the photovoltaic component contours, and generating a candidate contour set S; detecting corner points of all candidate contours, and calculating an included angle between an adjacent corner pointvector of each candidate contour and an edge of a photovoltaic component contour to which each candidate contour belongs; when the number of included angles meeting a predetermined angle condition issmaller than a preset threshold ratio, judging that cracks exist in the glass of the corresponding photovoltaic module. Through the invention, glass crack detection of the photovoltaic module can berealized automatically, accurately and timely.

Provided is a number identification method based on a SVM. The method comprises the following steps of: 1) acquiring a number plate image I1; 2) performing color preprocessing on the image I1 to obtain a grayscale image I4; 3) performing Canny operator edge detection on the image I4 to obtain an edge image I5; 4) performing digital area rectangle framing on the edge image I5 to obtain a contour set k<3> including numbers, recognizing the contour set k<3> including numbers by using a trained number classifier in combination with the original image I1 to obtain specific numbers. The method may substitute for conventional manual classification, increases marathon photograph classification speed, is good adaptive to the flexible material of a number plate and high in recognition rate.