144 results about "Integral projection" patented technology

The invention provides a laser machining method applicable for a complicated curved surface. By adopting processing steps of dividing the complicated curved surface into curved surface sheets and establishing a coordinate system of each curved surface sheet, the subsequent processing steps of dividing sub blocks and machining pattern parallel projection can be implemented according to the coordinates of the curved surface sheets respectively, so the limit that a machining molded surface backing to a laser incident direction or a steep surface parallel to an optical axis direction cannot be machined because only one coordinate system is adopted to perform whole injection in the prior art can be overcome, the machining efficiency is improved and the machining applicable range is extended. The invention also provides a laser machining device. The laser machining device has a three-axis laser machining head structure provided with a two-axis laser polarization mirror and a Z-axis mobile mechanism, so the problem that the conventional system depends on a three-axis lathe is solved; therefore, a machining unit is independent and compact, and is easy to assemble and disassemble; the conventional five-axis linkage milling machine can be changed into a laser-etching machining lathe so that the functions are compatible; and the laser machining device has a significant practical value.

The invention relates to a lip for an excavation bucket. The lip includes a lower plate defining a plurality of integral projections along a front edge thereof, an upper plate and a central plate sandwiched between the lower plate and the upper plate so as to form a laminated structure. The upper plate also defines a plurality of integral projections which together with the projections on the lower plate form tool attachment formations in a serrated front edge of the laminated structure for supporting ground-engaging tools on the lip.

The present invention is directed to provide an eye opening degree estimating apparatus capable of properly estimating the opening degree of an eye. In an eye opening degree estimating apparatus, a search axis used for estimating the eye opening degree is set in an eye region image. By integrating luminance values of the eye region image in positions in the direction of the search axis along the direction perpendicular to the search axis, a vertical-direction integral projection histogram is generated. On the basis of a feature amount of at least one of the vertical-direction integral projection histogram in a position in the direction of the search axis in which the vertical-direction integral projection histogram has an extreme value and the eye region image, the opening degree of an eye included in the eye region image is estimated.

The present technique provides for the generation of density maps using one or more basis material decomposition tables or functions. The basis material decomposition tables or functions are generated by simulating the system response to various lengths of basis materials using component characteristics of the CT system as well as the attenuation coefficient for the desired basis material. Measured projection data may be processed using the basis material decomposition tables or functions to provide a set of density line-integral projections that may be reconstructed to form a density map or image.

The invention discloses a face recognition-based fatigue driving detection method. The method includes the following steps that: human face video image acquisition is carried out; an integral projection algorithm is adopted to carry out eye and mouth positioning on the first frame of image, the images of a left eye portion, a right eye portion and a mouth portion are extracted; an adaptive particle filtering-based eye tracking algorithm is adopted to carry out eye tracking; an improved horizontal projection-based eye and mouth state recognition method is adopted to carry out recognition; eye and mouth fatigue information extraction is carried out, a PERCLOS value, blink frequency, eye closing time and the number of times of yawning are calculated; and fatigue judgment is performed on a driver based on the PERCLOS value, the blink frequency, the eye closing time and the number of times of yawning. According to the defects of an original horizontal projection method, an improved horizontal projection method is adopted to recognize eye and mouth states, and therefore, high recognition accuracy and high adaptability can be achieved; and the fatigue judgment algorithm can improve judgment accuracy.

The invention discloses a method for extracting a text in a complex background image. The method comprises the steps of firstly, using the susan operator to detect and identify angular points in a source image, after removing isolated angular points, conducting integral projection transformation to cut out suspected text regions, and screening and removing non-text regions according to priori knowledge; then, judging the background complexity of a text region by the utilization of gray level jump information, when a background is judged to be complex, conducting color clustering on the text region by the utilization of the kmeans clustering algorithm, and determining the type which a text belongs to and extracting the text according to the color information at the position where the angular points are the densest; when the background is judged to be simple, conducting binaryzation on the image by the utilization of a largest-between-class variance method; finally, realizing accurate extraction of the text region. The text extraction method can locate the text region in the complex background image, and finally extracts characters after removing the background.

A TAMR (Thermal Assisted Magnetic Recording) write head uses the energy of optical-laser excited surface plasmons in a scalable planar plasmon generator to locally heat a magnetic recording medium and reduce its coercivity and magnetic anisotropy. The planar plasmon generator is formed as a multi-layered structure in which one planar layer supports a propagating surface plasmon mode that is excited by evanescent coupling to an optical mode in an adjacent waveguide. A peg, which can be a free-standing element or an integral projection from one of the layers, is positioned between the ABS end of the generator and the surface of the recording medium, confines and concentrates the near field of the plasmon mode immediately around and beneath it.

Methods and systems for obtaining a motion vector between two frames of video image data are disclosed. Specifically, methods and systems of the present invention may be used to perform a block-matching algorithm over a two-dimensional search area in a manner that reduces number of comparisons. In particular, the method determines a best candidate block for each strip based by searching in a first dimension of a two-dimensional search area and based on a predetermined difference criterion. The method then determines a second set of best candidate blocks by performing a limited search in the other direction based on the results from the search in the first dimension. The method then determines a motion vector for the best candidate block. Integral projection arrays may be used to further optimize the search. The methods and systems of the present invention may be used in optimizing digital video encoders, decoders, and format converters.

The invention relates to a kilowatt-hour meter image automatic identification method which comprises the following steps: 1. image preprocessing: detecting vertical texture of a panel image by using Sobel operator, preliminarily removing the background area by a projection method, extracting the area with abundant vertical texture by an expansion method, and carrying out binarization treatment on the image by a adaptive threshold segmentation method based on an integral projection method; 2. precise positioning of indicating value and bar code: by combining an intelligent judgment method on the basis of indicating value intervals and length-width ratio characteristic of numeric characters under the complex image background, adapting to precise positioning of indicating values of different types of kilowatt-hour meters on the basis of vertical edge detection of the Sobel operator and morphological treatment; carrying out horizontal scanning on the bar code area to extract the bar code characteristic area; 3. bar code identification: identifying different character bar codes by using a similar edge distance normalization method; and 4. indicating value identification: extracting the indicating value by a PCA (principal component analysis) method. By using the PCA character recognition method, various character indicating values can be precisely identified, including identification of half-character.

The invention discloses a self-adapting motion compensation frame rate up-conversion method based on a motion classification mainly solving the problems of inaccurate motion compensation and over big motion complexity in the prior method and comprising: firstly dividing a video into a global motion sequence and a local motion sequence by a global motion estimating method based on an integral projection; obtaining a still background and an occlusion region by a motion detection to the local motion sequence through a high order statistical model; constructing a gray co-occurrence matrix with a pixel distance, quantizing a vein complexity by using two indexes of an angle second moment and an entropy; adding a Cross-Sectional Absolute Deviation (CSAD) criterion to a provided mixing motion estimation method, performing the motion estimation to a motion region of the global motion sequence and the local motion sequence according the vein complexity, avoiding the generation of a superposition and a pore zone during a motion compensation interpolation; finally performing the compensation interpolation to the motion region, and performing a time-domain average to the still region. The occlusion region is a copy of a current frame or former frame.

The invention discloses a human eye positioning method and device which are used for solving the problem that in the prior art, human eye positioning can not be conducted on a low-resolution face image. The invention further provides a human eye region positioning method and device which are used for solving the problem that in the prior art, human eye region positioning is poor in robustness with interference from glasses and eyebrows. The human eye region positioning method comprises the steps that a left eye region and a right eye region of a face grayscale image are determined; according to horizontal grayscale integral projection, vertical grayscale integral projection, horizontal gradient integral projection and vertical gradient integral projection of the obtained left eye region, the position of the center of a left eyeball in the left eye region is determined; according to horizontal grayscale integral projection, vertical grayscale integral projection, horizontal gradient integral projection and vertical gradient integral projection of the obtained right eye region, the position of the center of a right eyeball in the right eye region is determined.

The embodiment of the present invention provides a method, a server and a system for measuring a three-dimensional shape of a highly reflective surface. The optimal projection gray value is determinedby using a surface image sequence of a first measured object. A local overexposure problem easily occurring in the three-dimensional measurement process of the highly reflective surface is solved. According to the surface image of a second measured object and a phase unwrapping algorithm, coordinate matching between an image projection device and an image acquisition device can be carried out while reducing the overall projection intensity. An adaptive grating fringe image generated according to the optimal projection gray value has fewer projected fringe images; the operation is simpler; andthe calculation is less. Projecting to the surface of the measured object can prevent point cloud missing in an overexposed area during the three-dimensional reconstruction. Compared with the adaptive fringes generated by the traditional method, the modulation and contrast are better, and then the surface shape measurement accuracy and measurement speed of the measured object obtained by the three-dimensional reconstruction are ensured.

A threaded anchor which comprises a threaded shank portion, a tip portion defined upon one end of the threaded shank portion, and a head portion disposed adjacent to the opposite end of the threaded shank portion. The head portion comprises a hexagonally configured drive section for threadedly driving the threaded anchor into a substrate. An axially extending lug-type projection is integrally connected to the hexagonally configured drive section of the head portion, and a transversely oriented throughbore is defined within the axially extending lug-type projection so as to accommodate any one of a structural or fixture support wire or cable, a power cable, a fiber optic or telephone cable, a turnbuckle wire assembly, a support strap, a safety or guy wires, a tie down wire or cable, and the like.

The invention provides a motion parameter estimation method based on gray integral projection cross-correlation function characteristics. A horizontal moving motion model is adopted, and the method comprises the following steps that a target image is divided into N*N rectangular partitioning blocks with the equal size, gray integral projection is performed on each partitioning block, the variance of gray integral projection of each partitioning block is determined, R areas with the large variance can be selected as alternative areas, the cross-correlation operation is performed on gray integral projection of the alternative areas in two continuous frames of the target image, second derivatives of a cross-correlation function extreme point are calculated, weighted summation is performed on the variance of gray integral projection and the second derivatives of the cross-correlation function extreme point, good and poor factors of the alternative areas are obtained, the good and poor factors are arranged in a sort descending mode, the Q areas arranged in front are selected as the good-quality areas, local motion parameter estimation is performed on the obtained good-quality areas respectively, horizontal moving motion parameters in the direction of an axis X axis or an axis Y are obtained, and weighted summation is performed on the local motion parameters of the good-quality areas to obtain global motion parameters. According to the method, global motion parameters of an image with the low contrast ratio and the high resolution ratio can be effectively estimated.

The invention belongs to the machine vision and image processing field and specifically relates to a method for positioning human eyes in a human face image, wherein a polar-coordinate gray-level integral projection method is employed to perform human eye positioning. Polar coordinates are established with the central position of the human face as the pole, the horizontally leftward ray of the pole as the pole axis and the anti-clockwise direction as the positive direction of the angle. The gray level of the image is projected in the direction of the polar angle so as to determine the angles of the eyes; then, the pixels in the direction of the eyes are subjected to gray-level integral projection in the horizontal direction and the position of the minimal projection value is the projection position of the human eyes; therefore, the coordinate positions of the human eyes are determined and the human eye detection requirement of the inclined human face image is satisfied. The method for positioning the human eyes in the human face image can be applied to the human eye identification field.

A knuckle arm is connected to a machine body and a top cover of a machine, and includes a mounting seat, a connecting seat, a sliding seat, and a resilient member. The mounting seat includes a mounting block disposed fixedly on the machine body, and a pivot block extending integrally from the mounting block and having a cam edge. The connecting seat is connected pivotally to the pivot block of the mounting seat, and is attached to the top cover. The sliding seat is disposed slidably within the connecting seat, and has an integral projection. The resilient member is disposed between the mounting seat and the sliding seat, and biases the sliding seat to move toward the mounting seat so as to keep the projection of the sliding seat in contact with the cam edge of the mounting seat.

The invention discloses a facial feature extraction method, which comprises the following steps of: (1) dividing H, S, V (Hue, Saturation, Value) components of a facial image in an HSV color space and taking out the V component; (2) filtering the V component image; (3) corroding and expanding the obtained image; (4) carrying out binaryzation on the obtained image, and obtaining an image with black and white colors; (5) respectively carrying out integral projection on the obtained black and white image in a horizontal direction and a vertical direction to obtain a horizontal and vertical integral projection image; (6) determining feature values of primary organs of a face according to the integral projection image obtained through analysis; and (7) locating correct locations of the primaryorgans on the face according to the obtained feature values. Due to the adoption of the extraction method, accuracy rate of facial feature recognition can be improved.

A mounting device for supplying both high and low voltage services in an outdoor setting. One end of the mounting device is adapted to be anchored securely in the ground. A post extending from the ground includes a first electrical box for housing high voltage electrical connections and devices and a second electrical box for housing low voltage connections and devices. If desired by the installer, the mounting device can be configured to provide a single electrical box, which can include either high or low voltage. The top of the mounting device is adapted to accept a light fixture. The ground-anchored portion of the mounting device includes a large entryway to allow a large passageway for lead in of electrical cables. Integral projections extending from the mounting device enable secure anchoring to the ground. The mounting device is molded in one-piece and includes a knockout plate for providing a large electrical enclosure and a voltage separator for dividing the enclosure into separate high and low voltage electrical boxes.

The invention relates to an image processing and mode identifying technology, in particular to an orthogonal generalized projection-based analyzing method applied to document scanning, analyzing and processing. The analyzing method comprises the following steps of: (1) determination of an effective region in an document image: locating an effective ROI (Region of Interest) in the document and removing four margin regions in the document image to obtain a minimum rectangular region containing the ROI; (2) 90-degree deflection detection of the document image: determining whether the image is deflected for 90 degrees by calculating a generalized projection curve variance of the document image, and performing corresponding processing on the document image; and (3) 180-degree inversion detection of the document image: determining whether the image is inverted for 180 degrees by calculating a mean piecewise function of the vertical grey integral projection of the document image which is processed in the step (1), and performing corresponding processing on the document image.

The present invention relates to detection method of human face characteristics of motion image, firstly, take picture of human face image to construct discipline collection, then proceed principal component analysis, Hough conversion etc. to make its position and size coinside with the eye in discipline collection iamge, then project onto space of above mentioned characteristic eye, finally, take the eye being selected with minimum error between original eye and its projection as the detection result, and use integral projection to obtain precision position of mouth, nostril and apex nasi. Compared with existent method, the detection speed raises 225 times and accuracy increased by 1.27%.

The invention relates to a computer-vision-based system and method for detecting the number of pedestrians waiting to cross a crosswalk. The system comprises two charge coupled device (CCD) cameras and a computer image processing system, wherein the computer image processing system is connected with the cameras through interfaces; the two CCD cameras are arranged at two ends of the crosswalk; and field angles of the two CCD cameras cover an opposite pedestrian waiting region and a diagonally opposite vehicle region. The method comprises the following steps of: acquiring an image from a video image; preprocessing the pedestrian waiting region; acquiring a foreground image by using a Gaussian mixed background model method; acquiring a vertical integral projection graph; and processing the vertical integral projection graph, and performing information statistics to obtain the number of the pedestrians waiting to cross the crosswalk. Compared with a conventional pedestrian number detection method, the method has the advantages of simplicity and high efficiency, and can be applied to a complicated situation.

According to the invention, the human eyes are accurately positioned by a scale-invariant gradient integral projection algorithm and a Kalman filtering method, the degree of fatigue of the human eyes is detected by a calculation method based on the PERLOG feature, and different intervention measures are taken according to the degree of fatigue. The process mainly comprises the following steps: first, an image frame sequence of a user when the user uses a smart phone is collected through the front-facing camera of the smart phone, all the frames are sequentially processed by the smart phone, and the smallest rectangle containing the human eyes is positioned; second, the smallest rectangle containing the human eyes is positioned within a certain range of accuracy, the rectangle obtained each time is scaled according to the same fixed ratio, the area of the rectangle is worked out, the opening and closing condition of the eyes is judged according to an eye opening and closing threshold obtained through training, and the PERLOG feature value for judging the degree of fatigue of the human eyes is updated; and third, corresponding intervention measures are taken according to the degree of fatigue.

The invention provides a method for detecting a lorry safety belt hanging state based on road monitoring equipment. The method includes: step 1: obtaining a monitoring image from the road monitoring equipment; step 2: after preprocessing the monitoring image according to car and background object characteristics, adopting integral projection to locate a car body area in an original monitoring image; step 3: using hough conversion to locate a car window area in the car body area; and step 4: performing straight line detection; using a straight line detection method to detect whether straight line slope in the car window area is between 0.75 and 1.2, if the straight line slope is between 0.75 and 1.2, judging that a driver in the car window area belts up, and otherwise, judging that the driver in the car window area does not belt up. The car body and car window areas are utilized to locate the area where a safety belt is, thereby gradually diminishing a detection range, improving the accuracy rate and efficiency of safety belt detection, human eyes can be substituted and automatic judgment is performed, thereby improving the accuracy rate and efficiency of safety belt hanging detection.

A device for mounting a wide variety of different window dressings is described. The device is intended to use conventional mounting brackets to mount the window dressing without the necessity of providing holes to mount the same in the face of the window molding or frame. The device includes either a corner or linear mounting plate which, respectively, may be used either on the right hand or left hand corner, or horizontally or vertically on the frame. Each plate is U-shaped in cross-section and intended to cover the sides and front of the window frame or molding or L-shaped to cover a portion. In one embodiment the portion covering the face can be laterally adjustable. Another embodiment provides flanges for mounting directly on the wall surrounding the window when no molding is present. The front face of the plate may be provided with a plurality of mutually spaced dimensioned cut-outs. A universal clip which is a channel shaped member having integral projections extending from the back surface thereof may also provided. The projections are also mutually spaced and dimensioned to register on the cut-outs in the mounting plate and to be received therein to secure the clip to the mounting plate. The clip in turn receives a conventional bracket or an interface plate which is a track like member attached to either a shutter, or a conventional type bracket wherein the interface plate itself is slidably received in the clip. An alternate embodiment shows a projection extending outwardly from a mounting bracket on the molding to receive the foot of a tension rod. The bracket can cover an inside or outside portion of the molding and be mounted on the molding by one or more tacks. Another embodiment mounts a box for the ends of a mini-blind and slots or holes can be provided for axial pins of rollers for shades.

The invention discloses a human image optimization method capable of automatically detecting mask in human face key areas. The human image optimization method mainly includes two steps of human face detection and image smoothing, wherein the step of human face detection includes: using an adaboost algorithm, haar features and a cascade detector to detect a human face area, adopting a three-part five-hole rule to roughly divide a human face into several areas, judging positions of organs of the human face according to a binarization result of a picture, haar features and integral projection, and taking the organs which are detected as mask; using a principle of double-index edge smoothness maintaining filtering to perform image smoothing on a skin portion of the human face, and using the mask to cover the picture after being smoothed. The human image optimization method has high adaptive ability and reliability and is more accurate in organ detection and edge smoothness maintaining of the human face.

The invention relates to the technical field of image processing and computer vision and particularly relates to a solar cell broken gate defect detection method based on a densely connected convolutional neural network. The method comprises a step of dividing an acquired polysilicon solar cell broken gate defect image into a training set and a test set, a step of extracting an interested target candidate region by horizontal integral projection and training a two-class classifier based on the densely connected convolutional convolution neural network through the training set, a step of sending the test set into the trained classifier to classify and detecting a defect area, and a step of calculating a communication area and removing a detected discrete position according to the scale distribution characteristics of the broken gate defect and drawing a circumscribed rectangle of the defect position to obtain a detection result. According to the method, the structure of the densely connected convolutional convolution neural network is used to train the image block classifier for the first time to realize defect detection, the defect image can be accurately detected in a complex anddiverse background, the position of the defect area is given, and the automatic monitoring of product quality is completed.

A grid detector detects the existence and the location of grids in DCT compressed videos. When a grid is detected in the input video, a post-processor is turned on and the de-blocking processing is applied on the grid detected by the grid detector. When no grid is detected, indicating that the input video is either an uncompressed video or an already de-blocked video, post-processing turned off to avoid degrading the picture quality. To detect grids, the grid detector: (a) computes horizontal and vertical second derivatives for all pixels of the image; (b) generates horizontal second derivative zero-crossing mask and vertical second derivative zero-crossing mask by marking the those pixels whose second derivatives have opposite signs with respect to their horizontal or vertical neighboring pixels'; (c) applies horizontal and vertical integral projections to the horizontal and vertical zero-crossing masks respectively; (d) generates the local maximum masks by locating the local maximum of the two projected 1-D signals; and (e) determines grid location by computing the positions of the local maximum masks.

The invention discloses a medical image retrieval method based on deep learning and Radon conversion, and relates to the field of computer vision and image retrieval. In the crude retrieval stage, a BING target suggestion algorithm is adopted for detecting a region with a remarkable object, a partial mean value Pooling is introduced into deep convolution network architecture, region-based remarkable differentiation characteristics can be extracted, the characteristic dimensions are reduced, and then polymerization is performed to form global characteristic expression. In the characteristic vector quantization process, and a product quantization algorithm is used for solving the problem that characteristic vector similarity measurement calculation is high in complexity. In the fine retrieval stage, Radon conversion is used for performing integral projection on the image at multiple angles, more detailed information characteristics of the image can be obtained, and the Top 50 images obtained in the crude retrieval are subjected to Radon conversion to generate Radon bar codes, and more accurate retrieval is achieved through similarity measurement. The accuracy of the medical image retrieval is improved, and the medical image retrieval problems that the characteristic differentiation is not high and the characteristic dimension is high due to direct use of a convolution neural network are solved.