63 results about "Morphological transformation" patented technology

The invention discloses an image-based taillight detection and recognition method. The method comprises acquiring the real-time image of a vehicle ahead by using an ordinary camera and performing preprocessing by gradient sharpening and image cutting; performing adaptive threshold segmentation in combination with HSI and RGB color spaces and extracting the color information of taillights; extracting an outline by filtering denoising and morphological transformation, and constraining the same group of taillights by using a geometric condition; and hierarchically processing the state information based on a SVM and outputting the semantic explanation of the taillights of the vehicle ahead. The method, as an important link in an onboard advanced driving assistance system, has a good processing effect and a real-time processing ability for the detection of the taillights of the vehicle ahead and state information judgment in complex urban environment.

The invention discloses a positioning correction part in a QR code detection process, and a positioning and correction algorithm on the condition that one of three image seeking patterns of the QR code misses. The algorithm binarizes the image, then finds out the other two images besides the damaged image and calculates the version number of QR code. For version 1 QR codes, morphological transformation and edge extraction are carried out, and then the vertices are detected by line detection, and then corrected. For QR codes with version number 2 and above, the correction pattern of the lower right corner in the standard pattern of the QR code is searched, and the correction transformation is carried out by using the center points of the image seeking patterns and the center point of the correction pattern. The invention can accurately position and correct the QR codes on the condition that one of the image seeking patterns of the QR code misses and certain perspective distortion exists.

This invention deals with an integrated morphological system, herein called a morphological genome (morph genome), for design applications. This is the genome that encodes all form and is similar in intent to the biological genome that encodes all living things by its genetic code. The morph genome comprises a finite set of morphological genes, each gene specifies a distinct group of morphological transformations, each group of transformations is defined by a group of independent topological, geometric or other parameters. The morph genes and their parameters are mapped within an integrated higher-dimensional framework, with each parameter being represented along a vector in higher-dimensional Euclidean space. Each distinct number associated with a parameter or a group of parameters is represented by a distinct point in this space and is referenced by the higher-dimensional Cartesian co-ordinates of that point. This space uses a combination of discrete and continuous values for these parameters. The specific co-ordinates of any point in the space represent a genetic code for the specific form being mapped at that point. The entire space is a map of the morph genome and encodes all possible morphologies. The morph genome can be used as a design tool to generate known and new forms for applications in all fields of design including architecture, product design, environments and spaces, building or engineering structures, graphics, art, sculpture, technological devices, etc. The underlying method for the model for the morph genome applies to other fields of knowledge for systematically organizing and creating new concepts, structures, designs, information and taxonomies. The morph genome can also provide a basis for an integrated, interactive modeling environment for computers. It is a pedagogical tool for discovery and invention as well. Other applications include a graphic code for encryption as an alternative to numeric codes, and a system for representing numbers as integers, rational numbers and real numbers.

The invention relates to a license plate positioning method and a license plate positioning system based on cascading morphological transformation. The method comprises: performing downsampling and grey processing on an original image to obtain a gray image; performing cascading morphological transformation and binaryzation on the gray image, primarily screening a communicating area to obtain a plurality of candidate areas; computing the similarity between the color histogram information of each candidate area and the color histogram information of a standard license plate, and judging a license area according to the similarity; mapping the license plate areas into the original image through the inverse operation of downsampling, and thereby obtaining the corresponding original license plate area. According to the license plate positioning method and system based on cascading morphological transformation, gray processing, cascading morphological transformation and binaryzation are performed on the image to improve robustness in the positioning of candidate areas of the license plate and positioning speed of the license plate, meanwhile interference of strong light is suppressed, the contrast ratio of characters in the license plate area is increased, and screening accuracy of the license plate area is improved; the license plate positioning method and system based on cascading morphological transformation are applicable to environments with different light conditions and complex scenes, and can be used for simultaneously positioning the plurality of license plates.

The invention discloses a method and system for detecting road marker lines based on a geometric relation. The method and system extract each frame of image from a video, gray and binarize a color image, then perform morphological image processing, extract line features by using Hough transform, and combine discrete line segments by using the geometric relation to form road marking lines, correctsome abnormal line segments in view of a relation between frames, and finally draw the road marking lines in the frames and merge all the frames into a video. The method and system remove noise by morphological transformation, extract line features by using the Hough transform, combines the discrete line segments by the geometric relation to form the road marker lines, correct the abnormal line segments by using the inter-frame relation, are simple in algorithm, less in memory consumption and fast in processing speed. The method and system can be applied to road marker line detection in intelligent transportation.

The invention is suitable for the technical field of the city planning technology, and provides a quick extraction method and device for a city built-up area boundary, and terminal equipment. The method comprises the following steps that: obtaining the panchromatic image and the multispectral image of a city, and carrying out image fusion on the obtained panchromatic image and multispectral image;carrying out blocking processing on the fused image to obtain a plurality of multi-spectrum sub-block images, and adopting a multi-spectrum image edge detection operator to carry out edge extractionand image binaryzation on a single-spectrum component in the multi-spectrum sub-block images; carrying out morphology transformation processing on a binary image; carrying out Hough transformation onthe image subjected to morphological transformation to obtain the road of an image and carry out road elimination; and adopting the edge detection operator to obtain an image boundary line of which the road is eliminated to obtain a city built-up area boundary image. By use of the method, the image is subjected to edge extraction through sub-blocks, algorithm efficiency is improved, and meanwhile,the multi-spectrum image and the panchromatic image are fused to improve the accuracy of the algorithm.

The invention discloses an agaricus bisporus grading judgment method, and belongs to the field of agaricus bisporus production. The device comprises the following steps: S1, acquiring mushroom image information; S2, extracting a region of interest S3, segmenting images, wherein S31, converting a to-be-extracted image into a grayscale image; S32, adopting an OSTU threshold segmentation algorithm for processing; S33, carrying out morphological transformation; S34, performing expansion operation on a morphological transformation diagram; S35, transforming distance; S36, binarizing fixed threshold by taking the distance as a threshold to determine a foreground image; S37, subtracting the foreground image from the background image, and extracting an image contour; s38, finally obtaining the boundary of the original image according to the watershed change of the uncertain area in the markers; S4, calculating the number m of pixels of the foreground image according to the obtained foregroundimage; and S5, judging the grade according to a preset rule and the area of the mushroom pileus. The classification judgment accuracy of the agaricus bisporus can be improved, and errors caused by manual classification are reduced.

The invention relates to an aqueous phase preparation method for a single crystal silver nanosphere, and belongs to the technical field of nano materials. According to the main technical scheme, undera cationic surfactant solution system, firstly, a strong reducing agent is used for reducing 3nm gold nano seeds synthesized by chloroauric acid, then based on the seed-mediated growth method, a lessreductive reducing agent is used for sequentially growing the gold nano seeds to obtain 10nm gold nano seeds and silver nano cubes with different sizes, and finally, a sodium hypochlorite solution isadded to the silver nano cubes as an etching agent to induce the morphological transformation of the silver nano cubes. The prepared silver nano sphere has single crystal morphology, a regular shape,a uniform size and excellent tunable plasma resonance characteristics. The aqueous phase preparation method does not require complicated auxiliary equipment, the cost is low, and popularization is facilitated.

The invention relates to a bistable shell structure and a manufacturing method of the bistable shell structure. The bistable shell structure comprises a shell body and rib structures. The rib structures are buried into the shell body or protrude from at least one surface of the shell body. The manufacturing method of the bistable shell structure comprises the steps of rib structure formation and shell body formation, wherein the step of rib structure formation comprises embedded material formation or pre-formation conducted through a mold. The rigidity, strength and morphological transformation efficiency of the bistable shell structure can be effectively improved. Besides, because of the locally arranged rib structures, the weight of the bistable shell structure cannot be greatly increased.

The invention discloses a rotor rubbing acoustic emission signal denoising method based on a conjugate gradient method and adaptive generalized morphological filtering. The invention puts forward a generalized morphological filter which utilizes the conjugate gradient method to realize the generalized morphological open and close of the filter and combines adaptive weighting on the basis of basic morphological transformation and a combining form of the mathematical morphology, and the generalized morphological filter is applied to the denoising of a rubbing acoustic emission signal. The adaptive generalized morphological filter based on the conjugate gradient method, which is put forward by the invention, can be used for obtaining a better denoising effect on the rubbing acoustic emission signal and a higher signal-to-noise ratio compared with a traditional morphological filter.

The invention relates to an autonomous robot inspection method based on double yellow line detection. The method comprises steps that an image pre-processing step is carried out as follows, 1.1), image down-sampling and interest-area extraction are carried out; 1.2), the yellow information and the gray level of an image are extracted; 1.3), histogram equalization and binarization are carried out; and 1.4), morphological transformation is carried out; 2, zodiac line identification and center line positioning are carried out as follows, 2.1), refinement and Hof linear detection are carried out; and 2.2), intermediate line positioning of the double yellow lines based on a vector is carried out; 3, the mobile strategy is carried out as follows, a point D is searched from a straight line C, a y value of the point D is a fixed value, an x value changes with the straight line C, through determining the position of the point D, mobile robot decision is carried out. The autonomous robot inspection method based on double yellow line detection is advantaged in that accuracy, flexibility and economic property are realized.

The invention discloses a pistachio nut head and tail replacement and posture adjustment rotary mechanical device and method. The device comprises a computer, an industrial camera, an imaging structure and the like. The camera collects images, and the computer identifies the outline and inclination angle of pistachio nuts. The angle extraction algorithm comprises the steps of median filtering, blue component extraction, image segmentation, connectivity analysis, morphological transformation, edge detection, contour extraction, ellipse fitting and inclination angle extraction. The specific control method comprises the steps that (1) when pistachio nuts enter a head-tail replacement rotating station, comparing the current angle with a preset angle for judgment, and determining an angle adjustment value; 2) converting an adjustment value into a pulse signal and transmitting the pulse signal to a servo motor, and 3) outputting power of rotary motion by the servo motor, and driving the pistachio nuts to rotate through a coupler, a rotary main shaft and a rotary supporting platform in sequence. The inclination angle of the pistachio nuts is extracted by adopting intelligent vision, so that manual operation is effectively replaced, and the production efficiency is improved.

A physical process for recovering iron from magnetic cementitious spherical particles generated from metallurgical by-products such as slag from a basic oxygen converter (B.O.F.S) by applying a thermal process assisted by Ar—He plasma, which allows the microstructural and morphological transformation of these compounds based on iron, resulting in the generation of magnetic phases, among others. The separation process is based on the classification of the particles according to their new magnetic properties. Thus, when applying a magnetic field, the particles are separated into two fractions: a highly magnetic one and with low cementitious properties, and another fraction with high cementitious properties and almost no magnetic properties. The highly magnetic fraction is mixed with an aqueous medium to form an agglomerate and facilitate its transportation to be reefed to an iron melting process. The fraction with low magnetic properties has potential applications as a densifier in non-conventional cement and concrete.

The invention relates to a pressure alarm instrument positioning and reading identification method, which comprises the following steps of acquiring an original image, and extracting an instrument image through color identification, wherein theinstrument image comprises a background area and a barometer reading area, obtaining a barometer reading area by performing graphic morphological transformation and contour extraction on the instrument image, and obtaining the areas of a second green area, a second red area and a black area of the barometer reading area through calculation, and obtaining a reading result of the barometer by comparing the areas of the second green area, the second red area and the black area of the barometer reading area. According to the method, the barometer reading area is positioned and read through color identification, the problems of factors such as imaging noise, complex background, inclination, rotation and distortion can be eliminated, the target area is accurately positioned and classified, and the instrument reading is accurately identified for the target area.

The invention discloses a multi-modal small target detection method and system based on a remote sensing image, and the method comprises the steps: collecting data containing RGB images and infrared images, and making a data set; performing morphological transformation on the RGB image and the infrared image; constructing a multi-modal small target detection network model through a shallow feature network, a balance fusion network, a deep feature network, a spatial pyramid layer, a path aggregation network and a multi-scale head network; designing a loss function based on a multi-scale head network, and training and testing a multi-modal small target detection network model through the data set; and performing small target detection on a to-be-detected multi-modal remote sensing image through the tested multi-modal small target detection network model. The method can be suitable for small targets, and complementarity of the RGB image and the infrared image is fully utilized. In the multi-modal data based on the remote sensing image, the method has an accurate detection effect.

Disclosed is a water-in-oil-in-water (W/O/W) double emulsion including a first water phase, an oil phase and a second water phase, wherein the W/O/W double emulsion is disposed in an isotonic solution, and related methods of making the W/O/W double emulsion. Also disclosed is a method of making an artificial antigen presenting cell including: providing a W/O/W double emulsion that is stored in an isotonic solution, wherein the W/O/W double emulsion includes a peptide associated with a Major Histocompatibility (pMHC) complex or a glycolipid antigen associated with a CD1d molecule, and a costimulatory molecule; and transferring the W/O/W double emulsion to an electrolyte solution, wherein the double emulsion undergoes a morphological transformation to become the artificial antigen presenting cell. Also disclosed is a method of drug delivery including administering to a subject a unilamellar vesicle containing the drug. Other methods relate to storing a protein including making a water-in-oil-in-water (W/O/W) double emulsion, wherein the W/O/W double emulsion includes the protein, and wherein the W/O/W double emulsion is configured to be stably stored.

The present invention belongs to the field of medical rehabilitation training equipment, and particularly discloses a reversible mechanical arm gravitational torque balancing device, comprising a counterweight guide groove module, a counterweight, a mechanical arm joint, a rope, a guide pulley block, a counterweight disc, a mechanical arm joint link and a rack. The counterweight guide groove module is mounted on the rack; the counterweight has a lower end mounted on the counterweight guide groove module and an upper end connected to the rope, and the rope is wound on the counterweight disc after passing through the guide pulley block mounted on the rack; the mechanical arm joint is mounted on the rack and internally provided with a motor; the counterweight disc is mounted on an output shaft of the motor of the mechanical arm joint; and the mechanical arm joint link is also mounted on the output shaft of the motor of the mechanical arm joint and the gravity of the mechanical arm joint link produces a gravitational torque on the mechanical arm joint. The present invention can change the direction of the provided balance torque when the mechanical arm performs the morphological transformation, and features simple, rapid and labor-saving operation as well as reliable structure.

The invention discloses a novel three-rotor amphibious manned aircraft, and relates to the field of manned aircraft. The novel three-rotor amphibious manned aircraft is an amphibious vehicle capable of controlling deformation of a wheel wing and can be vertically taken-off and landed, a rechargeable storage battery is used for providing power supply, a servo system and a single-chip microcomputerare used for controlling and carrying out information processing, and the three-rotor amphibious manned aircraft can be normally operated through a control motor, a hydraulic system, light, an instrument panel and the like. The novel three-rotor amphibious manned aircraft has two operating modes involving a joystick and a holding handle, under a flight mode, an operating mode similar to an unmanned aerial vehicle joystick control is adopted, under a riding mode, an originally vertical grab handle is rotated outwards by 90 degrees along a steering wheel to become a transverse holding type handle. Through combined work of a hydraulic pump, a motor, a bevel gear and the like, the purpose of morphological transformation between flight and riding can be achieved.

The invention discloses a method for extracting image features based on a multi-film CT image and a 3DCNN network, and the method comprises the steps: carrying out the preprocessing of a pathological CT image, obtaining an original image, carrying out the morphological transformation of the original image, extracting a pathological first appearance image, filtering the original image through the pathological first appearance image, obtaining a qualified image, carrying out the image segmentation of the qualified image, extracting a second pathology profile image, performing mask operation on the original image by using the second pathology profile image to obtain a local pathology image, inputting the original image, the second pathology profile image and the local pathology image as three-channel input images into the 3DCNN network, and extracting pathology image features. According to the method, the pathological region is segmented from the CT image, multiple modes are formed by the original image, the pathological region profile image and the pathological region image, multiple groups of 3D images are generated, comprehensive analysis is carried out, and the judgment accuracy is improved.

The invention provides a computer-based Chinese character skeleton generation method, and the method comprises the following steps: a, visualizing a set representative character according to a configured standard font and a pixel area to obtain a picture representation form of the representative character; b, applying graying, binaryzation and morphological transformation to obtain a required representation word binary image representation form; c, splitting the Chinese character graph into primitives, independently storing the primitives, and refining the primitives into primitive skeletons by using a refining algorithm; d, recombining the skeleton primitives after affine transformation by using parameters in the database; e, cutting redundant pixel elements to obtain a skeleton of a newcharacter to be generated. According to the invention, the problems of long period, tedious work, low efficiency, need of a designer to work for a large amount of time and high cost due to the fact that the existing word stock needs to be designed one by one are solved.

A process for forming nanostructures comprising the step of applying on localised regions of a smooth thin film of bistable or multistable molecules an external perturbation with preset magnitude thereby said film undergoes a collective morphological transformation and nanostructures are formed by selforganisation of said molecules, said nanostructures having preset number, size, interspacing and shape. The nanostructures can be used as storage medium in storage devices.

An urban building material classification method fusing multi-source satellite remote sensing data comprises the following steps: S1, preprocessing a hyperspectral image and a high-resolution image; S2, carrying out independent component analysis on the preprocessed hyperspectral image to obtain an independent component, and carrying out matching analysis on the independent component and a building material to obtain a building material independent component image; S3, performing morphological transformation on the building material independent component images to obtain various building material feature images; S4, performing multi-scale segmentation on the preprocessed high-resolution image to obtain a building boundary image; S5, subjecting the building material feature image and the building boundary image to decision fusion to acquire a building material classification result. The hyperspectral image and the high-resolution image are fused to achieve building material classification, the accuracy of the classification result is improved, and the problems of geometric shape missing, boundary incompleteness and the like are solved.