53 results about "Geometric matching" patented technology

An automated registration and authentication system combines a generative and discriminative approach to improve the matching of a query object to a database of registered objects. The discriminative approach uses a voting mechanism to identify a most likely match, and the generative approach uses ASIFT transforms to determine a best geometric match. The two results are combined using a technique base on Bayesian inference theory.

The invention discloses a heat treatment method of a wheel rim surface of the train wheel with high carbon steel. The method is carried out as follows: during the cooling process, the wheel tread of train wheel is firstly cooled by spraying water with small flow in short time, then large flow in long time. The invention also discloses a heat treatment device of using the method. An water feed tubular tube comprises a water feed tubular tube with large flow and a water feed tubular tube with small flow, the water feed tubular tube with large flow is communicated with a plurality of nozzles with large flow, and the water feed tubular tube with small flow is communicated with a plurality of nozzles with small flow. Adopting the technical proposal can improve the contacting fatigue resistance of train wheel, reduces the tread spalling of train wheel; obtains the appropriate geometric matching relation between wheels and rails as soon as possible, increases the contact area, decreases the contact stress level; promotes the wear resistance of the internal part of wheel rim, thereby promoting the service life of wheel; can not increase the operation difficulty of the heat treatment and cooling technology of wheels.

The invention relates to a machine vision based liquid level sub-quality product marking method. The method includes: placing a certain qualified glass bottle at a work station to be tested, acquiring a template for subsequent matching treatment; conveying a glass bottle to be tested to a shooting work station through a conveying belt, and shooting an image of the liquid level of the glass bottle to be tested; performing binarization processing on the image after pretreatment; matching a bottle cap by using a geometric matching method, and detecting a position of the bottle cap; establishing a coordinate system; setting an area-of-interest ROI of a liquid level qualified area; matching a liquid level line by using the geometric method; and determining a sub-quality glass bottle. The machine vision based liquid level sub-quality product marking method is simple in algorithm, is fewer in steps, can achieve rapid detection of images, is high in detection efficiency, cannot be affected by environments, and is high in accuracy rate.

The invention provides a geometric matching method comprising steps of: training a matching model according to a template drawing image pyramid; searching a matching pose from the topmost image of theimage pyramid of an acquisition image, transferring the matching post to the next-layer image layer by layer and obtaining a high-precision matching pose, determining that the matching is successfulif the matching score of the matching pose in the bottommost image of the image pyramid is greater than a set threshold, otherwise determining that the matching fails. When the template drawing is trained, a certain number of probes are extracted from the image in the image pyramid of the template drawing. During the acquisition image searching or matching, the matching score is calculated by using the gradient information of the probes and the gradient information of the probes at the corresponding positions on the acquisition image. During the top searching of the image pyramid of the acquisition image, a formula of calculating the matching score is modified and a method of calculating the matching score is improved so as to improve search efficiency. During non-top matching, the translation, the scaling and the rotation in the pose are iterated and adjusted by many times to obtain the high-precision matching pose.

A first map comprising local features and 3D locations of the local features is generated, the local features comprising visible features in a current image and a corresponding set of covisible features. A second map comprising prior features and 3D locations of the prior features may be determined, where each prior feature: was first imaged at a time prior to the first imaging of any of the localfeatures, and lies within a threshold distance of at least one local feature. A first subset comprising previously imaged local features in the first map and a corresponding second subset of the prior features in the second map is determined by comparing the first and second maps, where each local feature in the first subset corresponds to a distinct prior feature in the second subset. A transformation mapping a subset of local features to a subset of prior features is determined.

The invention provides an area-array gazed infrared remote-sensing-image space-time three-dimensional noise identification and compensation method. In the method, aiming at an unstable three-dimensional noise generated by row, column and time axes in an area-array gazed infrared image, accurate noise positioning and compensation processing are realized. The method is realized by five processes ofsample selection, spatial filtering, noise primary selection, noise positioning file making and noise compensation. The method has characteristics that the method does not rely on blackbody calibration information, noise positioning is accurate, noise pixels are only processed and image information is not lost. When an existing noise filtering method is applied to an area-array infrared remote-sensing image sequence possessing the space-time three-dimensional unstable noise, a calculated amount is large, the image information is easy to lose and so on. In the invention, the above disadvantagesare overcome; the method can be applied to the on-orbit processing of an infrared image, quality improvement and other fields; and area-array infrared-image geometric matching accuracy can be increased.

The invention discloses a foamed aluminum prefabricated plate. The foamed aluminum prefabricated plate comprises a foamed aluminum prefabricated plate body, the periphery edges of the foamed aluminum prefabricated plate body are designed into a geometric structure in the thickness direction and matched with one another in a geometric mode, the matching mode of the opposite sides of the foamed aluminum prefabricated plate body is opposite to the geometric matching mode, and when the foamed aluminum prefabricated plate body is used in the geometric matching mode, the opposite sides of the foamed aluminum prefabricated plate body can be connected through one or two or more than two matching modes including rivet connecting, welding connecting, bolt connecting, adhesive connecting and the like. According to the foamed aluminum prefabricated plate, the structure is simple, and design is reasonable; when structural material connecting is conducted, the foamed aluminum prefabricated plate connected in the geometric matching mode is easy to install and fix and good in sealing property; the electromagnetic property of an aluminum-based alloy material can be significantly optimized and improved by changing the proportion of raw materials, and sounds, light, electromagnetic waves and the like can be effectively blocked.

The embodiment of the invention provides a house type code matching method and device, and relates to the technical field of home decoration design. The method comprises the following steps: receiving house type geometric data of a to-be-processed house type input by a user; performing rectangle decomposition on the house type geometric data to decompose a polygon corresponding to the house type geometric data into a plurality of rectangles; taking the center of the main rectangle with the largest area as an original point, and obtaining position information of each rectangle, a door and a window; encoding according to the position information to obtain a spatial code; and matching the space codes with existing space codes in a database to obtain the most similar house type, and carrying out coding and geometric matching on the single space of the house type, so that the similar space can be quickly screened out, and the problem that the matching effect of the existing method is relatively poor is solved.

The invention discloses a graph matching method with dynamic space regularization with a window, which belongs to the field of computer vision and pattern recognition, and more specifically relates to a graph matching method with dynamic space regularization with a window. According to the local feature description of the reference image and the image to be matched, the matching degree between nodes is measured by dynamic space regularization with a window, and the graph matching method is realized, which has the characteristics of less memory consumption and non-exact matching. Aiming at the ε-adjacency graph relationship between the node and its adjacent nodes in the reference image G1, determine the image G2 to establish the subgraph to be matched according to the number of nodes; use the cosine of the included angle in the subgraph to represent the local information of the central node, and use the windowed dynamic Space regularization is used to measure the degree of geometric matching between subgraphs and subgraphs, and the degree of matching support for the central node is given; then the node pair with the highest degree of matching support is selected to complete the graph matching process.

A method of updating navigational map data with missed attributes via polyline geometry matching includes obtaining first and second navigational map data sets. Chains are built of original road elements of the first navigational map data and chains are built of original road elements of the second navigational map data set. A spatial index of the chains is created. A set of chain candidates that are within a predetermined distance of each other is created according to the spatial index. The size of the set of chain candidates is reduced utilizing an angle filter. Each remaining chain candidate of the reduced set of chain candidates is ranked and the chain candidate having a best ranking is selected as the best candidate chain. The first navigational map data set is updated utilizing attributes from the best candidate chain.

The invention provides a large-scale face database searching method based on a shape space, and belongs to the fields of face recognition and shape analysis. In the large-scale face database searching method based on the shape space, face characteristic points are automatically located, an image is projected to the shape space based on locations of the face characteristic points, and a geometrical structure of the image for searching in the shape space is calculated. The method comprises two parts including an offline database creation part and an online searching part. The offline database stores the geometrical structure of the face image. The online part calculates the geometrical structure of the image to be searched, and quickly searches for a matched face image in the database. The large-scale face database searching method based on the shape space provides a novel design mode for the large-scale face database searching. The large-scale face database searching method based on the shape space is compact in match information data size, and is applicable to be implemented by a distributed system. The geometrical structure of the point set in the shape space can greatly reduce the matching number, accordingly increasing the searching speed, and being applicable to applications having higher requirements to instantaneity. The large-scale face database searching method based on the shape space can effectively solve the problems that the data storage size of the large-scale face database searching is enormous and the searching takes too much time.