93 results about "Thinning algorithm" patented technology

The invention relates to the technical field of photoacoustic imaging and discloses a rapid photoacoustic tomography device and a rapid photoacoustic tomography method based on limited-angle scanning. The method comprises the following steps of: transmitting pulse laser into an imaging sample to generate a photoacoustic signal, acquiring the photoacoustic signal at a limited position on an arc through an ultrasonic detector unit, amplifying the signal through a signal amplifier, converting the photoacoustic signal into an electric signal through an analog / digital (A / D) converter, and finally transmitting the electric signal to an oscilloscope to finish digital signal acquisition; inputting the photoacoustic signal of the imaging sample into a computer, performing filtering and Fourier transform processing on the photoacoustic signal through the computer, extracting the frequency domain information of the photoacoustic signal according to experiment conditions, establishing a forward projection operator and a measurement matrix, and performing reconstruction imaging on the signal based on a quick thinning algorithm through the computer. The method and the device have the advantages of short signal acquisition time, high reconstruction speed, convenience in operation, high adaptability, high expansibility and the like.

The invention provides a method used for automatically cutting X-typed overlapped and conglutinated chromosomes, comprising the steps as follows: step 1: binarization segmentation processing is carried out to a chromosome image so as to obtain a binarization template; step 2: the X-typed overlapped chromosome image is extracted from the template, the chromosome profile is obtained by a gradient operator and treated with Freeman chain coding so as to obtain a boundary chain code; step 3: the overlapped chromosome is processed by thinning algorithm so as to extract a single-pixel framework and determine the central point of an overlapping area; and step 4: the distance from the profile pixel point to the central point of the overlapping area after the Freeman chain code is calculated; four profile concave points which are not adjacent to each other are selected as cutting points so as to divide the profile into two profile sets; and the segmented chromosome profile sets are used for automatically completing the operation of segmenting the chromosome grey images, thus obtaining the automatic segmenting result of the X-typed overlapped and conglutinated chromosome. The method has quick and effective cutting operation and greatly reduces the working quantity and difficulty of manpower segmentation.

The invention relates to the technical field of image processing, in particular to a 3D skeleton modeling and hand detecting method. The method includes the steps that a figure video is shot through a depth video camera; face detection is carried out in the video; based on face depth information, a body silhouette is extracted; the body silhouette is verified, whether the silhouette is a body silhouette is judged, if yes, the next step is executed, and if not, the face detection step is returned to be executed; image filtering and smoothing are conducted; a skeleton line is extracted by using a detailing algorithm; the distance of body parts is calculated, and a probability distribution diagram is established; all the parts of a human body are recognized; precise joint points are obtained, and coordinates of hands in the 3D space are also obtained; all the joint points are connected to form a complete 3D human skeleton; the 3D human skeleton and the coordinates of the hands are output; the body silhouette is tracked, and information is provided for the next frame. According to the 3D skeleton modeling and hand detecting method, people can be detected rapidly under an existing technical condition, a human skeleton model is obtained, each joint point of people can be accurately positioned, the 3D skeleton is established, the arithmetic speed is high, the computation complexity is low, the 3D skeleton modeling and hand detecting method is suitable for various complicated backgrounds, and only 5ms is needed for each image frame.

In a water resource remote sensing survey, water body information is required to be extracted by using a middle-high-resolution image along with fine management of the water resource. However, since a high-resolution satellite image has fewer bands, the construction of an automatic water body extraction model is not facilitated, and further relatively fewer methods for effectively extracting water body information are available, and the error extraction rate and the extraction omission rate are higher. The invention constructs a new SPOT5 image-based automatic water body extraction method. According to the method, the polarization of the grey value of a water body is realized by constructing a calculation model capable of homogenously enhancing a water body index and multiple indexes based on remote sensing image background value research, and high-precision automatic extraction of the water body is realized on the basis by using improved automatic threshold selection algorithm, mathematical morphological filtering algorithm and thinning algorithm and the like. Experimental comparison shows that the influence of topographic shadowing, an inhabited place and the like can be eliminated to the greatest extent, a larger water body can be effectively extracted, and higher extraction precision on a finer water body is realized.

The embodiment of the invention discloses a method and a device for generating digital terrain model (DTM) data for an electronic map. DTM raster data are converted and segmented into multiple tile DTM raster data, the tile DTM raster data are converted into DTM vector data of a regular triangular mesh, a thinning algorithm is adopted to thin the DTM vector data of the regular triangular mesh, DTM vector data of an irregular triangular mesh are generated, and the DTM vector data of the irregular triangular mesh serve as DTM data corresponding to a large-scale electronic map; and as for each piece of tile DTM raster data, a 2.5-dimensional DTM image corresponding to the tile DTM raster data is generated, and the 2.5-dimensional DTM image serves as DTM data corresponding to a small-scale electronic map. Through the above mode, clear demonstration application requirements by the large-scale electronic map can be met, and quick rendering application requirements by the small-scale electronic map can also be met.

The invention discloses a line structured light center line extraction method, which belongs to the technical field of machine vision, and comprises the following steps of: carrying out a series of processing such as clipping, image graying, image enhancement, image denoising, image binarization, morphological opening and closing operation, image light bar region segmentation and the like on an image acquired by a CCD (Charge Coupled Device) industrial camera; performing refinement processing by adopting a refinement algorithm to obtain an image containing a single-pixel light bar center line;and improving a Steger algorithm. ;firstly, determining a region of interest and performing median filtering on the region; secondly, moving on an image row according to the determined constraint threshold and a 1 * 5 movable template, and finding out a rough center point; solving a Hessian matrix through the separability and the symmetry of a Gaussian function; and finally, carrying out Taylor secondary expansion to obtain sub-pixel-level center coordinates. The algorithm is good in connectivity, free of burrs, easy to operate, high in operation efficiency, high in extraction speed and highin precision. According to the invention, the real-time requirement of a visual detection system can be met.

The invention relates to a real-time self-investigation new path generating method of a vehicle-mounted navigation map, which is characterized by comprising the following steps in the vehicle running process: (1) arranging a signal point recorder which automatically records the information of a signal point after being started; (2) recording and refreshing the signal point in real time through the signal point recorder, and finishing a path increasing process; (3) judging the rationality of the signal point sequence of the increased path based on the time criterion and the distance criterion;and (4) carrying out real-time smooth processing on the signal point sequence of the increased path by utilizing a Kalman filtering algorithm and a point range thinning algorithm in embedded navigation equipment. The invention ensures the accurate recognition of the newly increased path of the map in the vehicle running process, reduces the influence on the shape of a section of path by drift or loss of GPS (global positioning system) signals, generates a relatively accurate signal point sequence, and establishes the foundation for real-time self-updating operation of the map.

The invention discloses a detecting and tracking method for lane lines of autonomous vehicle navigation. The method comprises firstly converting video frames into grayscale images, describing a tight degree of local connection of each point in the images through fuzzy connection degree and membership function, and extracting general contour of roads; secondly, extracting edges by using a Canny operator, refining the images by using a refinement algorithm and detecting control points of the road edges; and thirdly, fitting the edge points of the roads by using a non-uniform B-spline interpolation (NUBS), that is, calculating the fitting points by using controlling points. Experimental results show that the method takes grayscale of the roads and gradient characteristics into account, and thus reduces interferences of shadows or other noises and is relatively high in accuracy. A probability of successful detection of the lane lines is 92 % for detection of 50 images.

The invention relates to an intelligent positioning method facing a microfluidic chip, the use of chip planar images in CCD camera acquisition microfluidic chip analysis system, using morphological filters and OPTA (one-pass thinning algorithm) refinement algorithm to denoise and refine the chip ichnography, being a skeleton-map of the chip ichnography; based on chip skeleton map, extracts relevant node on the network of micro-channels of microfluidic chip; According to the joint position of all micro-channels and relations between the two connection, generates adjacent table; According to the adjacent table on the microfluidic chip for intelligent tracking positioning, makes the adjacent table feedback amendments through the relevant feedback algorithm in accordance with the positioning results. It is based on the relevant image processing technology designed with a fully automated positioning, high precision, high speed and can track position, and other characteristics, it can improve the automaticed analysis for all microfluidic chip analysis system.

The invention discloses a morphological image edge detecting method based on the quantum theory. The method comprises the following steps that S1, morphological edge detection is carried out based on the quantum theory; S2, edge thinning is carried out on an image in the mode of combining an NMS thinning algorithm and morphological thinning; S3, edge connection is carried out on the image through a quantum probabilistic method. When a quantum morphological edge detecting algorithm provided in the scheme is used for carrying out edge detection on a noise image, even though the strength of noise is large, the edge can be effectively extracted. Edge continuity can be effectively improved through an edge connecting algorithm based on quantum probability in the scheme.

The invention discloses a vehicle-mounted manhole cover detection method based on arc-point combination. The method comprises the following steps that images shot by an automobile data recorder are converted into grey-scale images, and an edge extraction algorithm and a thinning algorithm are used for contour extracting and thinning; according to connectivity and regional relationships among points, the contour is broken off from cross points, segmental arcs are classified according to concavity and convexity of each segmental arc, and linear segments and small segmental arcs are removed; candidate segmental arcs are divided into groups, arcs which meet quadrant constraint and central constraint are classified, and a central point position is determined by central point constraint; points are selected from segmental arcs in different quadrants and least square fit of an ellipse is carried out; false removal and screening are conducted on a fitted ellipse, and the position of a manhole cover is judged. According to the vehicle-mounted manhole cover detection method based on the arc-point combination, segmental arc processing is used as preprocessing of the point fitting, most of the irrelevant interference is removed at the complicated road background, the probability of invalid random sampling is excluded, and the detection speed of road surface manhole covers is improved.

The invention relates to a practical binary document image tilt angle detection method, which belongs to the technical field of image preprocessing. The detection method comprises: extending foreground pixels by traversing the foreground pixels in images to strengthen the characteristics of text lines, and then carrying out filtering smoothing; extracting edge characteristics by thinning algorithm, and then marking the pixels after thinning; counting up the number of pixel points, and filtering and binarizing the edge characteristics of the number of successive pixel points less than the threshold value; and finally, carrying out the Hough transformation for the images after processing, and detecting the tilt angle of the binary document images. The experiment proves that the invention has higher practical value.

The invention provides an automatic construction method and system for a road network of open mine. The method comprises: obtaining pre-acquired GPS point data of the road network of the open mine; according to a transverse span Xspan and a longitudinal span Yspan of the GPS data distributed in space and the precision Tol of the road network, constructing M*N grids; initializing a binary image with a pixel of M*N, and according to the number of GPS points located in each of the M*N grids, assigning a pixel point corresponding to the grid in the binary image with a value of 0 or 1; by using a Hilditch thinning algorithm, extracting a framework of the assigned binary image subjected to value assignment, and obtaining the grid road network of the open mine; and by using a grid image vectorization method, vectorizing the grid road network of the open mine to obtain a topological relationship network of an open mine road. The automatic construction method for the road network of the open mine can realize automatic construction of the road network of the open mine.

The invention discloses a ship historical track rule extraction method based on unsupervised clustering. The invention belongs to ship historical track rule extraction. According to the method, the clustering result is provided for sea area grid division and related statistical calculation, the purposes of noise track elimination and track pre-classification are achieved, and the final track abnormity detection precision is improved. According to the method, firstly, track points are compressed through a track thinning algorithm, then track line segments are clustered through a density-based spatial clustering algorithm (DBSCAN); on one hand, outlier abnormal track line segments are removed, and on the other hand, track line segments similar in position, course, speed and the like are combined into one class, so that separation of air routes is achieved. Finally, a de-noising and aggregated track cluster is formed for subsequent sea area grid division. The method solves the problem that in sea area grid training, the statistical result is influenced by the fact that routes cannot be distinguished and channels are staggered, and improves the accuracy of anomaly detection.

The invention discloses a subway tunnel segment dislocation quantity detection method. The subway tunnel segment dislocation quantity detection method includes the following steps of: inputting a three-dimensional data matrix of a depth image acquired by a Kinect device, and performing pre-treating on data; converting the depth image into a binary image capable of being processed by a data image technology, through a double diagonal difference algorithm; processing noise in the binary image through a combined de-noising method, and meanwhile, removing special noises from bolt holes and injected holes based on shape features; extracting a dislocation framework though a thinning algorithm, identifying different dislocation lines through a global search algorithm, and finding dislocation positions in the corresponding depth image to calculating the segment dislocation quantity. According to the subway tunnel segment dislocation quantity detection method, the calculation is simple, the operation time is short, and manpower is not required; the detection of the segment dislocation quantity can be completed only through inputting the acquired depth image of a subway tunnel segment by means of surface measurement, thus, the detection method is high in efficiency and is accurate in detection.

In the three-dimensional reconstruction process based on mesh projection, the mesh refinement is a crucial step and has a significant impact on the subsequent point matching. In order to solve the mesh refinement problem, the invention first adopts a refinement algorithm based on a mathematical morphology and improved by an index table to perform initial refinement processing on a binocular vision image. Then, for the problem that a frame line is depressed towards the internal line in the operation process of the refinement algorithm, the invention successfully solves the problem through an improved refinement algorithm based on linear fitting. The experimental results show that the algorithm improves the speed of mesh refinement and solves the problem of mesh depression.

The invention discloses a color image edge detection method based on fusion color gradient to solve the technical problem that the existing color image edge detection method is low in practicability. According to the technical scheme, the method comprises the following steps: first, a color image is transformed from a RGB color space to an HSV color space, the gradient of a hue component is calculated with use of circular hue distance, and the gradient of the hue component is calculated with use of symbol type color distance; then, a first principal component of the RGB color image is acquired by a PCA (principal component analysis) method, wherein the principal component contains richer edge information than a brightness image acquired by a traditional method; and the gradient of the first principal component is calculated with a classic gradient operator, the hue component and the gradient image of the first principal component are fused, and the final color image edge is obtained by a morphological thinning algorithm. By adopting the color image edge detection method, the defect that the gradient operator can be hardly applied to the hue component in the background art is overcome, and complete color image edge is obtained. The method is highly practical.

The invention relates to the field of image processing algorithms. According to the technical scheme, the famous high-quality tea picking point position information obtaining method based on machine vision comprises the following steps: 1, obtaining a tea image from a tea garden, and performing Gaussian filtering noise removal on the tea image through a 3 * 3 convolution kernel; 2, setting a respective ROI for each tender shoot obtained through the image; 3) converting the ROI in the RGB color space into an HSV color space, and extracting the features of the tender shoots and the growing point branches of the tender shoots; 4) carrying out secondary binarization segmentation on the extracted tender shoot and branch areas by using an Otsu algorithm; 5) refining the binarized picture in the previous step by adopting an improved Zhang refining algorithm, and extracting a skeleton of the binarized picture; 6) using a Shii-Tomasi algorithm to search bifurcation points of the tender shoots and the branches as feature angular points for detecting the refined skeleton; 7) fitting the lowest point and the angular point of the contour into a linear segment. According to the method, the precision and efficiency of tea tender shoot picking point positioning can be improved.

The invention relates to the field of multimedia information processing, and specifically to a video subtitle thinning method based on template matching and curve fitting. According to the video subtitle thinning method, taking structural features of Chinese characters into consideration, strokes of Chinese characters are extracted and redrawn, thereby achieving an effect of thinning the Chinese characters, and solving the problems of stroke distortion and loss and low recognition rate which are generated when other thinning algorithms are used to perform thinning on the objects of the class. The technical points of the video subtitle thinning method in the invention includes: obtaining Chinese character images in video subtitles; determining connected regions of the Chinese character images; proposing a template matching method to extract strokes of Chinese characters in the same connected region; and proposing a method of redrawing horizontal strokes and vertical strokes with horizontal and vertical line segments, redrawing left-falling strokes and right-falling strokes with secondary Bezier curves, and redrawing dot strokes with oblique line segments, thereby realizing a thinning effect of the Chinese character images. The video subtitle thinning method based on template matching and curve fitting in the invention keeps the shapes and the stroke structures of the Chinese characters, thereby preventing a distortion phenomenon; ensures flatness of horizontal strokes, straightness of vertical strokes and bending of left-falling and right-falling strokes, thereby better redrawing the shapes of the Chinese characters; and improves a recognition effect of the video subtitles.

The invention discloses a fast and effective fingerprint thinning method, which relates to a fingerprint thinning method and aims at solving the problems that the thinning is incomplete, a thinned ridge is not a single pixel, and the correctness in fingerprint recognition can be influenced in the existing fingerprint thinning method. Fast thinning: the judgment is first carried out to boundary points of fingerprint ridges in a picture to be thinned; the boundary points of the fingerprint ridges are gradually deleted to obtain a fast thinning picture. OPTA thinning: a removing template and a preserving template in the OPTA thinning method are utilized for changing the fast thinning picture into a single pixel picture. The invention adopts the combination of the fast thinning algorithm and the OPTA algorithm so as to significantly improve the operational efficiency of the thinning algorithm and solve the problems that a previously-thinned fingerprint picture has more burrs and branch points are incompletely thinned. Meanwhile, the invention can also ensure that the connection of the fingerprint ridges is not damaged; detailed characteristics of the fingerprint are protected; the arithmetic speed is relatively fast; a processed fingerprint picture can be completely thinned and the skeleton can be close to the center line of the fingerprint ridge.

The invention discloses a method and a device for mapping a river course map to a Web map. The method comprises the following steps: water area images of a river course map and a Web river course mapare respectively extracted; the topological structure of water surface is obtained by a thinning algorithm; according to the surface topology map, the pixel positions of the feature points and the GPSestimated positions are extracted, and the pixel positions of the feature points and the GPS estimated positions are stored in the corresponding nodes of the surface topology map; the pixel positionsof the feature points and the GPS estimated positions of the feature points are extracted from the surface topology map; a mapping point pair is obtained through a corresponding node; the transformation matrix of perspective projection from river map to Web river map and the origin of GPS mapping after transformation are calculated by using mapping point pairs; a perspective projection transformation matrix is used to complete the mapping from a river map to a Web river map. By adopting the invention, the accurate registration of the real geographical environment and the Web map can be realized.

The invention discloses an unlocking controller of a fingerprint lock and a multi-person series unlocking method. The unlocking controller comprises a fingerprint lock controller and a fingerprint sensor. The unlocking controller is characterized in that a light emitting diode for indicating the work state of the fingerprint lock controller is arranged on the fingerprint lock controller, a protective cover capable of awakening the fingerprint lock controller is arranged on the fingerprint sensor, a programmer and a fingerprint collector are also arranged in the unlocking controller, the fingerprint collector sends the fingerprint information of all relevant personnel to the programmer, and the programmer sets the fingerprint information for unlocking the fingerprint lock and storing the fingerprint information into the fingerprint lock controller. The invention changes the traditional method of storing the fingerprint of all relevant personal into the fingerprint lock controller, and greatly improves the confidentiality of the fingerprint. Because the unlocking sequence of the unlocking personnel can be limited, the fingerprint lock is safer and more reliable. Because a template matched pulse coupling neural network thinning algorithm is adopted, and the invention is realized on a DSP control platform, the fingerprint identification accuracy and the identification efficiency are improved.

The invention discloses a rapid implementation method aiming at an OPTA (One-Pass Thinning Algorithm) of an image, belonging to the technical field of digital image processing. The method mainly comprises two parts: constructing an equivalent lookup table according to a delete template and a reservation template in the OPTA; rapidly thinning a binary image according to the constructed lookup table. The invention has the advantages that the complicated template matching operation in the OPTA is converted into a completely equivalent rapid lookup operation, thereby greatly improving the operation efficiency of the OPTA.

The invention provides a method used for automatically cutting X-typed overlapped and conglutinated chromosomes, comprising the steps as follows: step 1: binarization segmentation processing is carried out to a chromosome image so as to obtain a binarization template; step 2: the X-typed overlapped chromosome image is extracted from the template, the chromosome profile is obtained by a gradient operator and treated with Freeman chain coding so as to obtain a boundary chain code; step 3: the overlapped chromosome is processed by thinning algorithm so as to extract a single-pixel framework and determine the central point of an overlapping area; and step 4: the distance from the profile pixel point to the central point of the overlapping area after the Freeman chain code is calculated; four profile concave points which are not adjacent to each other are selected as cutting points so as to divide the profile into two profile sets; and the segmented chromosome profile sets are used for automatically completing the operation of segmenting the chromosome grey images, thus obtaining the automatic segmenting result of the X-typed overlapped and conglutinated chromosome. The method has quick and effective cutting operation and greatly reduces the working quantity and difficulty of manpower segmentation.

The embodiment of the invention provides a method for realizing pulmonary nodule adaptive matching based on CT image skeleton registration. The method comprises the following steps of data preparation, lung and skeleton point cloud data extraction, three-dimensional point cloud data registration and pulmonary nodule adaptive matching. . Firstly, lung image data and pulmonary nodule data are subjected to three-dimensional point cloud rigid transformation registration based on the characteristic of small human skeleton change characteristics, so that alignment of the lung and pulmonary nodule data before and after follow-up visit is realized. Secondly, an FGR algorithm is adopted, so that the method is obviously superior to ICP and other local refinement algorithms in the aspects of operation speed and registration accuracy. Thirdly, RMSE is adopted as a lung point cloud registration error, adaptive matching of pulmonary nodules is achieved, manual intervention of pulmonary nodule registration is little, the automation degree is high, and the registration result is accurate; and fourthly, normalization processing is carried out on the CT image data, so that the robustness of the algorithm can be improved, and the method can be widely applied to different types of CT equipment and DICOM data with different pixel spacing values.

The invention discloses a grape vine winter shearing point detection algorithm, and belongs to the technical field of target detection. The method comprises the following steps: step 1, segmenting a grape vine area; step 2, refining and vectorizing a grape vine area; step 3, reconstructing a grape vine connection relation; step 4, finding grape vine branches needing pruning; step 5, determining bud points are detected, and pruning points. According to the invention, the SLIC superpixel method is adopted to segment the grape vine region and the background region in the grape vine image, and single pixel points are gathered into pixel blocks, so that the segmentation speed is greatly increased; color information features and texture information features are used as segmentation basis to train a BPNN classifier, so that the segmentation accuracy is improved; a refinement algorithm is adopted, so that the refinement time is shortened, the refinement precision is improved, redundant pixel points are reduced, and the single-pixelation of a refinement result is realized; according to the invention, the branch and trunk connection relation of the refined grape vine skeleton is reconstructed, redundant detection of other grape vine branches which do not need pruning is effectively avoided, and the efficiency is improved.

The invention discloses a blood vessel structure image plane atlas construction method, which comprises the following steps: 1) collecting a two-dimensional or three-dimensional medical image containing a local blood vessel structure, and if the original image is a blood vessel two-dimensional projection image, distinguishing and labeling arteries and veins; 2) extracting a blood vessel region byusing an image segmentation algorithm, and constructing blood vessel three-dimensional volume data; 3) extracting a skeleton line of the blood vessel region by using a thinning algorithm; 4) finding out a branch point and an end point of the skeleton line; 5) constructing a triangular curved surface of the blood vessel; 6) closing end points of the triangular network to form a discrete curved surface; and 7) mapping the discrete curved surface to a plane to form a discrete curved surface, and keeping all sides of the triangle not to be overlapped to obtain a planar two-dimensional image of thevascular structure. According to the invention, the three-dimensional structure of the blood vessel can be expanded to a two-dimensional plane on the premise of keeping the connection relationship between the blood vessels and the length information of the blood vessel, so that the hemodynamic characteristics of a local blood vessel area can be observed and analyzed conveniently.

The invention relates to a map image thinning and linear tracking method based on bit arithmetic, belonging to the field of geographic information identification and extraction. The method comprises the following steps: A, adopting a method based on adaptive thresholds to carry out the binarization treatment of map images; B, adopting a full eight-neighborhood connecting thinning algorithm based on the bit arithmetic to carry out the thinning treatment of the binary map images; and C, finally, sequentially carrying out linear tracking and data compression (removing redundant points) on the binary map images after the thinning treatment and storing result data. The new high-efficiency map image thinning and linear tracking method based on the bit arithmetic can be widely applied to the fields, such as the identification and the vector quantization of line symbols of the scanned map images, and the like.

The invention is a method for retrieving ground digital elevation from a single-flight full-polarization synthetic aperture radar image. Using the fully polarized scattering Mueller matrix solution, the migration of the orientation angle ψ is derived, and with the help of the transformation of the Euler angle, the connection between the migration of ψ determined by the three Stokes parameters and the range angle β and horizontal azimuth angle γ of the inclined surface pixel is established . Under the condition of only single-flight SAR data, adaptive threshold binarization and image morphology thinning algorithm are used to determine γ and β, as well as the ground surface range azimuth slope and horizontal azimuth slope. Therefore, the Poisson equation of the DEM is numerically solved with the complete multigrid algorithm, and the DEM inversion is obtained.