92 results about "Edge density" patented technology

A vehicle license plate recognition method and a system thereof are disclosed. A region where a vehicle license plate image exists is detected according to the edge densities of an input image and a vehicle license plate specification. A text area of the vehicle license plate image is divided into a plurality of character images. The character images are binarized to obtain a plurality of binarized character images. A plurality of characters is recognized from the binarized character images. The characters are recombined to form a character string. The abovementioned steps are repeated to obtain a new character string from another image of the same vehicle, which is captured at a next time point. The character string is compared with the new character string character by character to obtain a comparison result for verifying reliability of recognition through a voting technique.

A computer-interfaced camera system identifies and tracks groups of socially interrelated people. The system can be used, for example, to track people as they wait in a checkout line or at a service counter. In a preferred implementation, each recorded camera frame is segmented into foreground regions containing several people. The foreground regions are further segmented into individuals using temporal segmentation analysis. Once an individual person is detected, an appearance model based on color and edge density in conjunction with a mean-shift tracker is used to recover the person's trajectory. Groups of people are determined by analyzing inter-person distances over time.

The invention provides a method for screening images without mosaic or with little mosaic, and a system thereof; a method for screening clear images and a system thereof; and a method for screening images and a system thereof. The method for screening video images comprises the following steps of: converting the video images into gray scale images, filtering monochrome images; screening the images without mosaic or with little mosaic by the method for screening the images without mosaic or with little mosaic; calculating information entropy of the screened images in order to screen the images with more information quantity; dividing golden visual regions of the screened images in order to calculate marginal density of each region and obtain the marginal density ratio of the images; respectively calculating normalized marginal gradient values of the screened images; and calculating the weighting sum of the marginal density ratio and the marginal gradient values, and screening the video images with the maximum weighting sum. The method and the system provided by the invention can conveniently and fast screen the video images satisfying the needs, improve the screening efficiency, be good for searching and screening the video files, and improve loyalty of users to access the website.

A vehicle restraint system has a vision-based occupant classification system for control of airbag deployment during a crash scenario. The classification system utilizes two imaging sensors which together create a stream of paired images received and stored by an occupant classification controller. A computer program product of the controller utilizes the paired images to extract disparity / range features and stereo-vision differential edge density features. Moreover, the controller extracts wavelet features from one of the two paired images. All three features or maps are classified amongst preferably seven classifications by algorithms of the computer program product producing class confidence data fed to a sensor fusion engine of the controller for processing and output of an airbag control signal input into a restraint controller of the vehicle restraint system.

A computer implemented method, data processing system, and computer program product for establishing document relevance by semantic network density. When a search query is received, one or more semantic networks are identified which contain nodes matching one or more terms in the search query. An edge density is determined for each node matching a term in the search query. A relevancy score is then calculated for each of the one or more semantic networks based on the edge densities of the nodes matching a term in the search query. Based on the relevancy score, the relevancy to the search query of a document associated with the one or more semantic networks may then be determined.

The invention discloses a license plate positioning method based on multiple characteristics. The method comprises: the vertical direction edge density characteristic of an RGB colored image is utilized to detect the license plate area in an image and a binary image template in line with the license plate texture density characteristic area is obtained; the R, G and B pigment component value of each pixel of the RGB colored image are compared in an RGB colored space and a plurality of mask layer images of the pixels in line with preset color conditions are obtained; a logical calculation is carried out between each mask layer image and the binary image template, overlap areas are eliminated according to a preset color priority sequence and the mask layer image of each zero lap area is obtained; the abnormal area, the abnormal width high and width high ratio area, the abnormal edge density area and the abnormal color area of each mask layer image are eliminated to obtain a final license plate positioning image. The invention also discloses a corresponding license plate positioning apparatus based on multiple characteristics. The invention can be free from complex background interferences and is characterized by quick positioning speed, high accuracy and strong robustness.

The invention provides a dense population estimation method and a system based on multi-feature fusion. The method comprises the following steps: partitioning an image into N equal sub-blocks; performing hierarchical background modeling on the image by using a method based on a CSLBP (Center-Symmetric Local Binary Pattern) histogram texture model and mixture Gaussian background modeling, extracting the foreground area of each sub-block subjected to perspective correction, detecting the edge density of each sub-block in combination with an improved Sobel edge detection operator, and extracting four important texture feature vectors in different directions for describing image texture features in combination with CSLBP transform and a gray-level co-occurrence matrix; performing dimension reduction processing on the extracted population foreground partition feature vectors and texture feature vectors through main component analysis; inputting the dimension-reduced feature vectors into an input layer of a nerve network model, and acquiring the population estimation of each sub-block through an output layer; adding to obtain the total population. The dense population estimation method and system have high accuracy and high robustness, and a good effect is achieved in the population counting experiment of subway station monitoring videos.

The invention provides a video image character detecting method based on sparse expression, which comprises the following steps of: S1, resampling a video sequence to obtain a color video image, and converting the gray level and the multi-scale of the color video image to obtain a multi-scale gray level image; S2, performing edge detection and morphological closed operation to the multi-scale gray level image with an improved Sobel operator to obtain an edge image and filter the edge density of the edge image; obtaining a candidate character region through connected domain analysis and regular analysis; and S3, performing vertical projection and horizontal projection to the candidate character region, diving a vertical projecting image and a horizontal projecting image to obtain candidate character lines, dividing the candidate character lines into small regions through sliding windows, extracting the edge characteristics of the small regions, respectively classifying each small region with a classifying method based on the sparse expression, judging whether the small regions are character regions, judging the candidate character lines according to the judging result of the small regions to obtain and output a final character line region.

A surface wave plasma (SWP) source couples microwave (MW) energy into a processing chamber through, for example, a radial line slot antenna, to result in a low mean electron energy (Te). An ICP source, is provided between the SWP source and the substrate and is energized at a low power, less than 100 watts for 300 mm wafers, for example, at about 25 watts. The ICP source couples energy through a peripheral electric dipole coil to reduce capacitive coupling.

The invention provides a vehicle feature recognition device based on public security video images in skynet engineering. The device is characterized in that an image acquisition model is connected with a licence plate recognition model and a vehicle type recognition model respectively and is used for providing video frame images of moving vehicles; and the licence plate recognition model and the vehicle type recognition model are connected to a vehicle feature database server and store the feature parameters of licence plate information and vehicle type information recognized from the video frame images of the moving vehicles. The device has the following technical effects: videos are acquired in the places with complex public security such as main traffic thoroughfares, public security accesses, public gathering places, hotels and the like via the skynet engineering; video stream extraction frame by frame, licence plate positioning, licence plate segmentation, character recognition, moving vehicle outline extraction, vehicle edge density computation and decision tree based vehicle type recognition are carried out on the obtained public security video images, thus recognizing the feature parameters of licence plate information and vehicle type information; and the device has positive significance to further expansion and extension of urban public security prevention and control and urban comprehensive management tools.

The invention relates to an automatic extracting method of an urban built-up area border, and in particular relates to a method applied to a high-resolution panchromatic remote-sensing image and used for automatically extracting the urban built-up area border, and belongs to the field of urban planning. The method comprises the following steps: firstly performing histogram mean value stretching treatment on the image, and extracting a reasonable edge density area of the urban built-up area based on an edge detection operator, and extracting a textural feature reasonable area based on a gray level co-occurrence matrix and neutral network algorithm, operating the reasonable area with the edge density and the reasonable area of the textural feature to obtain a candidate area of the urban built-up area, finely extracting by using a partitioning density filter method and a mathematical morphology method, and performing smooth treatment on the border by using the mathematical morphology method and extracting the urban built-up area border. The region with a plurality of centers and relatively large space water areas can be effectively treated to make up the insufficiency of an existing method, and the extracted urban built-up border has the high precision.

The invention provides a water-free bridge target identification method in a remote sensing image, mainly solving the identification problem of a water-free bridge. The identification steps are as follows: (1) carrying out edge extracting for an original image by using Canny operator, and calculating the marginal density of the whole image and the marginal density of each pixel according to the given definition; (2) carrying out two-valued division for the original image by using the marginal density; (3) masking the original image with binary image; (4) carrying out edge extracting for the masked image by using the Canny operator, and converting and extracting straight line with Hough; (5) calculating the complexity of line segment according to the given definition, and finally determining the suspected bridge area; (6) respectively calculating the five textural characteristic quantities of suspected bridge area of the original image and the image after smoothened such as entropy, energy, correlation, local smoothness and moment of inertia so as to form a set of characteristic vectors; and the water-free bridge target is identified by using BP network to carry out classification criterion. The invention can be used for the water-free bridge target identification of the remote sensing image.

The invention belongs to the field of computer vision, and discloses a fixation point detection method based on local evaluation and global optimization. The method comprises the steps of extracting possible candidate targets in an image; then utilizing a supervised learning method to conduct local evaluation on the targets, wherein two evaluation methods are as follows, (1), grading significance of each proposals using image training SVM of a whole database; (2), utilizing a semi-coupled dictionary learning algorithm, reconstituting different SVM aiming at different images, and grading the proposals of the image with pertinence; after local evaluation, utilizing a proposal subset optimization algorithm to cluster proposals; finally conducting global optimization. According to fixation point detection method based on local evaluation and global optimization, aiming at the characteristics of different information catching the attention of human eyes, a model capable of seizing the information is designed, and the human eye gazing area in an image containing semantic information, an image containing objects, and an image which is complicated or does not contain the objects can be effectively detected.

A vehicle restraint system has a vision-based occupant classification system for control of airbag deployment during a crash scenario. The classification system utilizes two imaging sensors which together create a stream of paired images received and stored by an occupant classification controller. A computer program product of the controller utilizes the paired images to extract disparity / range features and stereo-vision differential edge density features. Moreover, the controller extracts wavelet features from one of the two paired images. All three features or maps are classified amongst preferably seven classifications by algorithms of the computer program product producing class confidence data fed to a sensor fusion engine of the controller for processing and output of an airbag control signal input into a restraint controller of the vehicle restraint system.

The present invention discloses a product multidimensional correlation-oriented degradation failure modeling method. The method specifically comprises: determining reliability characteristic quantities of a product; collecting and processing degradation data of each characteristic quantity; establishing a degradation model of each characteristic quantity; estimating a parameter in the degradation model of each characteristic quantity; establishing a multidimensional correlation degradation model by using D-Vine Copula; and determining types and parameter values of all bivariate Copula functions. According to the method provided by the present invention, a multidimensional correlation variable joint density function and an edge density function thereof are connected by a plurality of bivariate Copula functions by using D-Vine Copula, so as to effectively process a multidimensional degradation process with complex correlation, and achieve better flexibility and generalization ability.

The invention relates to the technical field of image processing, in particular to a method for measuring an average size of complicated particles based on Canny boundary detection. The method comprises the following steps of 1, inputting an image of moving particles; 2, performing Gaussian smoothing noise removal; 3, when the image size is greater than an image size threshold value, shrinking the image; 4, performing Canny boundary scanning, and converting the image into a binary image; 5, calculating the particle edge density of the image, and determining an average shape parameter of the particles and a background space parameter of the image; and 6, calculating the average size of the detected particles and the number of detected particles. through adoption of the method, the average size and the number of the particles in the image can be detected quickly and accurately; and the method is very suitable to statistics and measurement of a plurality of complicated particles at real time on line and can be used for pre-partitioning acomplicated particle image.

The invention discloses a mammary gland molybdenum target X-ray image block feature extraction method based on tower-shaped principal component analysis (PCA). The mammary gland molybdenum target X-ray image block feature extraction method based on tower-shaped PCA mainly overcomes the defect that features extracted in the prior art do not contain the feature that the density of the middle of a lump is large while the density of the edge of the lump is small. The method comprises the following steps of (1) carrying out pretreatment, (2) constituting a tower-shaped structure, (3) obtaining a gray feature vector of each image layer, (4) training a feature space of the gray feature of each image layer, (5) obtaining principal component features of each image layer, and (6) obtaining mammary gland molybdenum target X-ray image block features based on tower-shaped PCA. According to the method, the mammary gland molybdenum target X-ray image block features can be represented more robustly, image features can be represented more effectively, the accurate rate of detection of a lump region in a mammary gland molybdenum target X-ray photography image is increased, and therefore radiologists are assisted to carry out clinical diagnosis.

Disclosed is a density evolution algorithm based on a refined definition of node and edge densities for different parts of the code. In particular, density functions ƒV(1)(i) and ƒV(2)(i) of the output edges of the variable nodes with degree i within different codeword regions w1 and we, respectively, are defined and then calculated. Further, density functions ƒC(1)(j) and ƒC(2)(j) of the output edges for check nodes with degree j within codeword regions w1 and we, respectively, are defined and then calculated. Mixture density functions of output check edges connecting variable nodes in the first codeword region and the second codeword region are then calculated to determine an LDPC code design.

The embodiment of the invention provides a license plate detection method and system. The method includes the steps of obtaining a license plate image, carrying out gray-scale treatment on the license plate image to obtain a gray-scale image, carrying out discrete cosine transformation, sign conversion, square operation, inverse discrete cosine transform and binarization processing sequentially on the gray-scale image to obtain an edge obvious feature graph, searching for the area with the edge density larger than a preset threshold in the edge obvious feature graph to obtain an uncertain license plate area, and detecting the uncertain license plate area by means of a license plate classifier to obtain a license plate area. The method and system are more suitable for hardware implementation, and an algorithm can be transplanted into an embedded system more easily.

The present invention is applicable to the field of image processing, and provides a detection method of license plate and a device thereof, the method includes the following steps of: performing multi-scale detection for the images by using a HOG model and extracting a pane area via the HOG model detection; calculating an edge of the image and an integrogram image of the image after edge computation to gain the edge integrogram iamge of the image; gaining the pane area via the HOG model detection on the edge integrogram iamge, calculating an edge density of the pane area via the HOG model detection, and extracting the edge density greater than or equal to the pane of the threshold value; via a Haar model, detecting the edge density greater than or equal to the pane of the threshold value to gain the exact license plate area. The present invention by the HOG features detects the images, and then performs the detection the edge and the edge density, finally the license plate area is detected by the haar model.

The invention discloses an image fusion method combining wavelet and beamlet transform, and overcomes the defect of the fuzziness of image edges and detail positions of a wavelet-based image fusion method in the prior art. The method is implemented by the following steps of: (1) inputting a source image; (2) performing wavelet decomposition; (3) performing low-frequency sub-band coefficient fusion; (4) extracting an image edge; (5) calculating edge density and a difference value; (6) performing high-frequency coefficient fusion; (7) performing consistency verification; (8) reconstructing an image; and (8) outputting a fused image. By the method, the edge information of the image is extracted effectively by utilizing beamlet transform, and the edge information is participated into the image fusion, the interference of a noise to a fusion result is reduced, and finally the fusion image with more obvious edge detail features is obtained.

The invention discloses a method and device for identifying icons. The method comprises the following steps of: acquiring foreground edge density characteristics of images to be identified; and classifying: classifying whether the images to be identified are icons or not according to the foreground edge density characteristics. The method disclosed by the invention can be realized by a corresponding icon identifying device.

Disclosed is a method for creating a template for the purpose of performing pattern matching on the basis of a template image having high contrast. Also disclosed is an image processing apparatus. In the method for creating the template, design data is partially extracted, and on the basis of the extracted partial region, the template for template matching is created. In the image processing apparatus, such method is performed. In the method and the apparatus, a density of edges that belong to a predetermined region in the design data equivalent to the region to be searched for in the template matching is obtained.

A sample changer for transferring radioactive samples between a hot cell (12) and a measuring apparatus (10), as e.g. a hybrid K-edge densitometer, comprises a transfer channel (28) axially extending through a tubular containment (22) between the charging / discharging port and the measuring window section D. A recipient (24), with at least one compartment (100) for receiving therein a radioactive sample, is arranged in the transfer channel (28) so as to be movable therethrough. A threaded spindle (44) is rotatably housed in a spindle channel (42) below the transfer channel (28). A coupling (46) passing in a sealed manner through the closed rear end section connects a stepping motor (34) to the threaded spindle (44). A longitudinally guided support carriage supports the magazine (24) and engages the threaded spindle (44), so as to be subjected to a translational movement upon rotation of the spindle (44).

The present invention discloses a vehicle body color recognition region positioning method and device. The method includes the following steps that: the edge texture image of the vehicle head region of a vehicle body is acquired, and the integral image of the edge texture image is calculated; the integral image is divided into a plurality of local regions of first preset area, wherein the local regions do not intersect with each other; the unit area edge densities of the local regions are calculated, a first preset number of local regions are selected according to the unit area edge densities so as to be adopted as first regions to be recognized; local sub-regions of second preset area are selected from each first region to be recognized so as to be adopted as second regions to be recognized, wherein the local sub-regions have the smallest unit area edge density; and a second preset number of regions are selected from the second regions to be recognized according to the unit area edge densities of the local regions and the local sub-regions so as to be adopted as vehicle body color recognition regions. Since the regions with smooth textures are selected according to the unit area edge densities, and then the vehicle body color recognition regions are selected, and therefore, external interference resistance is excellent.

The invention discloses a town information extraction method based on a Google Earth remote sensing image. The method is characterized by, to begin with, performing pretreatment, such as median filtering and the like, on the remote sensing image; then, generating an edge image of the image based on a sobel edge detection operator; next, constructing a Gauss spatial domain weighting edge density generation operator and generating an edge density image; carrying out OSTU optimal threshold segmentation on the edge density image; and at last, carrying out aftertreatment on the segmentation image, and finally obtaining town information of the remote sensing image. The extraction method in the invention is good in effect, fast in speed, and high in accuracy, and is suitable for information extraction of towns, especially towns at the connection places of cities and suburbs.

In an image reading system, a photographic film is conveyed, an image frame is conveyed to a reading position, pre-scanned and fine scanned. An image obtained by pre-scanning is displayed on a display section. The film is conveyed after the inspection processing ends. Further, by using data output from six image surface detecting sensors disposed along a transverse direction of the film, inter-frame region candidates (edge candidates) are detected within an inspection range per frame. A region surmised to be between frames is detected from film light amounts input from the sensors. Edge slopes, the edges' density differences, are computed as evaluation values for detecting an edge to be used as a reference from among the candidates, and are compared. An edge having a largest density difference is extracted. With the extracted edge as a reference, an image frame position is determined while considering an image frame size.

The invention discloses a method for identifying video smog in combination with color and appearance characteristics. The method has good realtimeness and accuracy. The method first detects a foreground target of a current input frame by using a ViBe foreground detection algorithm, analyzes and trains, after morphological treatment, correlative data and computes the color probabilities of four channels including RGBa in order to remove a pseudo smog area, and eliminates other pseudo smog areas by appearance characteristics including smog and contour complexity, edge density and the like so as to further increase identification accuracy. A plenty of experiments and simulations manifest the good performance of the method for identifying video smog in combination with color and appearance characteristics.

The invention discloses a detecting method and a device of a vehicle logo. The method comprises that a region can be confirmed as a license plate region on a vehicle grayscale image, wherein the region can meet edge density reserved requirements after edge segmentation operation; a vehicle logo region is positioned with using license plate information and the license plate region and the vehicle logo is detected by using edge detection and watershed segmentation in the vehicle logo region. The accuracy of detecting the vehicle logo is improved through the detecting method and the device of the vehicle logo.

The present invention discloses a multi-station radar station location and joint tracking method based on the distributed PHD (pulse height discrimination). The method comprises the following steps: S1, receiving echo signals, and performing local tracking filtering processing; S2, calculating a Chernoff information divergence formula between posterior in every two radars; S3, constructing an optimization problem model; S4, solving the optimization model, and obtaining position parameters of all radars relative to other radar sites; S5, selecting multi-sensor information fusion criteria; S6, jointing the posterior distribution to change into an edge density function, and obtaining a fused posterior density function according to the multi-sensor information fusion criteria; and S7, transmitting the fused posterior density function back to the local radar in the form of mixed Gaussian. According to the method provided by the present invention, information of the target can be measured bymultiple radars in the case of the unknown multi-station radar precise position, multi-station radar location, multi-target tracking and information fusion can be performed, and the method has a mallcalculation amount, a fast convergence speed, and other characteristics.