151results about How to "Real-time extraction" patented technology

The invention discloses a visual online image ferrography imaging system, which comprises an image acquisition device, a lens base, a focusing mechanism, an LED (Light Emitting Diode) annular reflection light source and a bicolourable ultra-thin LED backlight source; the image acquisition device is sleeved and fixed at the upper part of the focusing mechanism through the lens base; the bottom of the focusing mechanism is connected with an oil flow channel in ferrography through a thread; the LED annular reflection light source is fixedly connected with a threaded lens cone in the focusing mechanism; the bicolourable ultra-thin LED backlight source is adhered to a sedimentary slide in the oil flow channel in the ferrography. According to the invention, the object distance and the image distance are respectively adjusted through the focusing mechanism; the focal distance error of the lens is corrected; ferrography imaging with the fixed rate of 1.8 times can be realized; simultaneously, high-brightness and uniform illumination can be provided by adopting the LED annular reflection light source; the bicolourable ultra-thin LED backlight source as the transmission light source is wide in illumination range; the full-field ferrography imaging can be realized; the visual online image ferrography imaging system disclosed by the invention has the characteristics of being small in volume, high in imaging quality and easy for extracting visual abrasive particle characteristic information.

The invention discloses a method for extracting a tower target in an unmanned aerial vehicle routing inspection power transmission line image, and belongs to the technical field of pattern recognition and intelligent computation technologies. The method for extracting the tower target comprises the steps of converting an image acquired by aerial photographing from an RGB color space into an HSI (hue / saturation / luminance) space, carrying out binarization processing on an I component image of the HIS space after being subjected to median filtering so as to acquire a binary image of an electric power tower profile; carrying out angular point detection on the binary image of the electric power tower profile by adopting an improved Susan angular point detection method, and marking positions of all angular points in the binary image of the electric power tower profile; and detecting the mass center of an angular point sequence based on a moment invariant mass center detection algorithm, and marking the position of the electric power tower in an original image by taking the mass center as a center. Through applying the technical scheme disclosed by the invention, the electric power tower can be extracted accurately in real time, and a guarantee is provided for fault diagnosis of electric power components such as insulators, hardware fittings and the like on the tower.

The invention provides a quick dynamic reactive-power compensating method of combined electrical load. The method is characterized in that a power electronic power converter (12) is controlled by an intelligent controller (7) to change the current and the impedance of a secondary winding of a variable reactor (11), so as to change the current and the impedance of a primary winding of the variable reactor (11); the fundamental current is changed by changing the admittance of the variable reactor when the input voltage is unchanged, so that the change of the variable reactor absorbing reactive power is caused; when an adjustable capacitor bank (5) is controlled by the intelligent controller and the target power factor is compensated, an inductive reactance convertor (6) can slide to adjust the reactive power in cooperation with the adjustable capacitor bank; therefore, the dynamic reactive-power compensation of the combined electrical load is realized. The method can realize real-time analysis and extraction of reactive power and reactive current under different load conditions, so as to realize intelligent control of the optimal quick reactive-power compensation; and the method can be used for quick reactive-power compensation of the combined high-voltage electrical load in the industries such as metallurgy, building materials, petrifaction and the like.

The invention discloses a phase delay extraction and compensation method for PGC phase demodulation. A sinusoidal phase modulation interferometric signal is converted into a digital interferometric signal through an analog-to-digital converter after amplifying and filtering, first-order, second-order and fourth-order harmonics of the digital interferometric signal are subjected to orthogonal downmixing simultaneously, and three pairs of orthogonal harmonic amplitude signals are obtained. Phase delays are extracted from the three pairs of orthogonal harmonic amplitude signals, and corresponding phase delay correction coefficients are calculated out by using the phase delays. The phase delay correction coefficients are multiplied by the absolute harmonic amplitude signal equal to the sum ofthe absolute values of the orthogonal harmonic amplitude signals to obtain new harmonic amplitude signals which are not affected by the phase delays, and then a to-be-measured phase is obtained through antitangent operation. The problems are solved that in the PGC phase demodulation technology, phase delay is difficult to measure and non-linear errors caused by the phase delay are difficult to compensate for in real time, the phase measurement accuracy is improved, and the method is widely applied to the technical field of sinusoidal modulation interference.

The invention discloses a road surface line laser rut detection and identification method and a processing system. The detection and identification method includes the steps that road surface rut laser line images are collected; characteristic extraction is carried out on the images to obtain a ridge characteristic strength graph; in the characteristic strength graph, numerical values of each row of pixel grey levels in the images present a waveform curve, one or more waveform peak points exist in each curve, the waveform peak points are candidate points of points on a laser line rut and called ridge points, and waveform ridge points of each row of the pixel grey levels are continuously looked for one by one according to peak values to determine a rut laser line. According to the method, the rut laser line can be automatically and precisely extracted in real time, the processing system has high robustness and provides precise and reliable data for calculation of subsequent road surface rut models and rut depths, and therefore measurement accuracy of the road surface rut depths is improved.

The invention discloses a static hand gesture description method based on the quadrangle skeleton structure; the skin-color region is detected to obtain the binary mask of the static hand gesture region; then the contour-rectangle is constructed according to the binary mask and is equally divided into four sub rectangles; the centroids of the binary mask pixel point sets in the four sub rectangles are searched in the four sub rectangles; the four centroids are connected to construct a quadrangle skeleton used for indicating different hand gestures; then six perimeter ratios, four angle perimeter ratios and four rectangle degrees of the quadrangle skeleton are extracted and used as the description characteristic values; the description characteristic values are processed normalization to form the 14-dimensional characteristic vector for expressing different static hand gestures. The static hand gesture description method not only can recognize the characteristics of the common hand gestures, but also can recognize the characteristics of the complex static hand gestures, so as to satisfy the real-time processing requirement and improve the recognition correct rate of the complex hand gestures.

The invention discloses a fetus electrocardiosignal extracting method based on a self-adaptation FLANN filter, and belongs to the technical field of fetus electrocardiosignal detecting. The fetus electrocardiosignal extracting method solves the problem that due to the nonlinearity of a mother electrocardiosignal from the chest area to the abdomen area and the non-stability characteristic of the electrocardiosignal , the fetus electrocardiosignal is extracted inaccurately. A mother and fetus mixed electrocardiosignal is obtained through the lead of the abdomen of the mother, and the mother electrocardiosignal is obtained through the lead of the chest of the mother; two kinds of electrocardiosignals obtained from the body surface of the mother are respectively processed; the preprocessed mother and fetus mixed electrocardiosignal and the preprocessed mother electrocardiosignal are normalized to serve as an original input signal and a reference input signal of the self-adaptation FLANN filter; parameters of the self-adaptation FLANN filter are constantly updated through the self-adaptation LMS algorithm, and the non-linear relationship of the mother electrocardiosignal from the chest area of the mother and the abdomen area of the mother is estimated; an error output signal of the self-adaptation FLANN filter serves as an estimated fetus electrocardiosignal. The fetus electrocardiosignal extracting method is used for extracting the fetus electrocardiosignal.

The invention provides a dynamic self-adaption skin color segmentation method and a device. The method comprises the following steps: a red green blue (RGB) image is obtained; colors and space of the RGB image are transformed to obtain a hue saturation value (HSV) image; a face of a person is tested to obtain a position and an area of the face; color values of a skin color pixels in the face area are collected to serve as a skin color sample, and color values of positions with intense light changes are recorded to serve as a skin color change sample; each chromatic value or saturation value of histograms of the skin color sample and the skin color change sample are polled and are combined with predetermined weight; through the covering rate of the pixel counts within the chrominance or saturation threshold ranges of the histograms to the total counts of the corresponding histograms, skin color segmentation threshold values are obtained; the skin colors are tested through the skin color segmentation threshold values. By means of the above method and device, the skin color dynamic and self-adaption segmentation can be achieved, and thereby the accuracy of the skin color testing and the adaptability of the skin color testing to light changes are improved.

The invention discloses a method for rapidly positioning a license plate in a video image. By the method, license plate candidate regions are rapidly segmented by utilizing a gradient shadow detection method and a one-dimensional filtering technology on the basis of vehicle bottom shadow characteristics and then are positioned precisely. The method comprises the following steps of: (1) extracting a frame of image in video and segmenting processing regions; (2) graying the images of the processing regions and performing gray stretch; (3) performing primary coarse positioning according to the vehicle bottom shadow characteristics; (4) de-noising by a one-dimensional filtering method and determining candidate regions; and (5) precisely positioning the candidate regions by using morphology. The method has the advantages of easiness in implementation, high robustness, high precision, high real-time performance and the like, and can provide a new real-time and reliable rapid positioning method for the license plate positioning in license plate identification.

The invention discloses a human body sweat real-time monitoring and sensing system. The system comprises a flexible extensible sensor chip array attached to human skin, a plurality of sensing chips are arranged, the plurality of sensing chips are connected in parallel through a flexible circuit, each sensing chip comprises a nano gold working electrode, an Ag / AgCl reference electrode and a multilayer graphene carbon nanotube microporous pipe connecting the working electrode with the reference electrode, the reactant is attached to the inner wall of the microporous pipe, the reactant corresponds to the constituents detected by each sensing chip, the multilayer graphene carbon nanotube microporous pipe has ductility and does not affect the movement of the skin when attached to the skin, andthe sweat reacts with the reactants on the inner wall of the pipe when flowing through the microporous pipe to generate electric current or voltage signals and determine the condition of each component in the sweat according to the monitoring signals to determine the human health conditions so as to perform early warning for possibly generated diseases and provide medical-grade information for clinical diagnosis and treatment.

The technical scheme provides a real-time digital video watermarking method based on an H.264 standard, comprising the following steps of: embedding watermarking information obtained after processing and coding and used as secondary enhancing information into a secondary enhancing information (SEI) field of each frame by analyzing the H.264 protocol standard and using available information resources in the H.264 standard; and ensuring robustness and nonvisibility of watermarking on the condition that video quality and a code stream format are not influenced. The real-time digital video watermarking method based on the H.264 standard can effectively resist frame deletion attack, has a simple realization mode, is used for solving various problems in the prior art that spatial synchronous distortion and photometric distortion can not be effectively resisted. The technical scheme can be used for judging copyright ownership or hiding, secretly carrying and transmitting information, and the like.

The invention discloses an AVS video digital watermark method based on nonnegative matrix decomposition, which mainly solves the problems of low transmittance and poor attacking repellence to frame deletion and frame recomposition. The invention generates watermark based on content by utilizing light characteristics of a video I frame and extracts kinetic characteristics of video signals by utilizing the nonnegative matrix decomposition during the watermark embedding, and can self-adaptively adjust the embedding strength and the embedding position of the watermark by combining texture and light characteristics of the video to embed the watermark into intermediate frequency DCT coefficients of prediction residual errors of an I frame light block. During the watermark verification, the I frame is only decoded, the video product is judged whether the watermark is embedded or not by comparison of the extracted watermark and the verification of the error rate of the watermark. The invention has high transmittance and robustness, can be used in the video product with AVS standard, and provides reliable copyright protection for the AVS video product.

The invention provides an SMT material mistake preventing and tracing method and system. The method comprises the following steps that a database is established based on material information, and management is conducted; bar codes are established for materials and stored in the database; the material bar codes are recognized, and matching is conducted on the recognized information and the information in the database; if the material bar code information is matched with the information in the database successfully, the next step is conducted; if the material bar code information is failed to be matched with the information in the database, an alarm is sent out; and the bar codes of the materials which are processed completely are scanned, the scanning information is stored in the database, and the purpose of extracting production line data in real time can be achieved. It is not required to make feeding tables, feeding is carried out according to programs, material mistaking is prevented, and paper-free operation is achieved. Manual checking is not required, the efficiency is high, and no mistake will be made. The materials are sought quickly, the line exchanging efficiency is improved, and the employee pressure is lowered.

The invention discloses a network flow analysis and file extraction system. The system comprises a flow capture module, a flow preprocessing module, a protocol statistics module, a protocol analysis module and a file extraction module. Under the condition that normal communication flow of an enterprise is not influenced, communication flow between the enterprise and a public network can be captured, flow of different IPs and different protocols is counted, transmitted files are extracted in real time, and file content is written into a disk.

The embodiment of the invention belongs to the technical field of signal processing and provides a method and device for extracting the main melody of music signals and terminal equipment. The methodcomprises the steps that: to-be-processed music signals are obtained, wherein the music signals comprise a plurality of frames of audio data; each frame of audio data is preprocessed, so that valid frames and invalid frames can be distinguished from each other; the invalid frames are identified as fundamental-frequency-free frames; for the valid frames, a frame frequency spectrum is calculated, and a main frequency is determined according to the frame frequency spectrum; a plurality of candidate fundamental frequencies of the valid frames are determined according to the main frequency of the valid frames and a preset harmonic energy cross-correlation algorithm; a fundamental frequency is identified from the plurality of candidate fundamental frequencies of the valid frames according to a preset rule; and based on the fundamental frequency of all the invalid frames, a fundamental frequency contour line matched with the music signals, i.e., the main melody of the music signals, is generated. With the method and device provided by the embodiments of the invention adopted, the main melody of the music signals can be accurately extracted in real time, and a calculated amount is small. The method can be used in various kinds of hardware containing microprocessors.

The invention discloses a real-time operation state collection system of a high-voltage mutual inductor group. The system comprises a synchronization module, a collection module, a merging module andan output module; the synchronization module comprises a Beidou / GPS reception module and a synchronization control core module, receives Beidou / GPS signals, provides PPS pulses and a time-service signal, and supports real-time synchronous collection among multiple collection systems; the collection module realizes A / D sampling and characteristic quantity extraction on secondary output electrical analogy signals of groups of mutual inductors; the merging module merges characteristic quantity extraction results of multiple channels, and extracts absolution time from the Beidou / GPS time-service signal; and the output module transmits a collection result to local and remote terminals, and the result can be used to evaluate an operation error of the mutual inductor group in real time. Via the system, operation state data of the groups of mutual inductors in the power grid can be collected synchronously in high precision and in real time, and data basis is provided for evaluation of the operation error of the mutual inductor group.

The invention relates to a field moving object fining extraction method, which comprises the following steps that: a sequence image including a moving object is obtained, and the sequence image is preprocessed; the preprocessed sequence image is subjected to frame-by-frame differentiation, and is then segmented into a plurality of grids, a located moving region of the object is determined according to feature values of the grids, and a grid method is used for extracting a moving region of the object for further reducing an object range; the background is subjected to modeling in the moving region of the object, a binaryzation chart of the object is obtained through a background subtraction method, and the binaryzation chart is subjected to feedback pixel processing; the processed binaryzation chart is mapped to a located color chart region of the object, and the color chart region is subjected to super pixel segmentation; and the segmentation result is fused with the binaryzation chart, the confidence degree of each super pixel is calculated according to the fusion result, and after the thresholding, the fining moving object is finally obtained. The field moving object fining extraction method can realize the real-time robust and fine field moving object extraction under complicated backgrounds.

The invention, which belongs to the technical field of electrical variable measurement, provides an adaptive-notch-filter-based harmonic wave extraction method. The method comprises: a, collecting a to-be-measured signal by a sampling device to obtain an input signal u(t); and b, carrying out harmonic wave extraction on the input signal u(t) so as to obtain each harmonic instantaneous value . / xh, a fundamental frequency omega 1, a direct current deviation A0, and each harmonic amplitude Ah. With the method, a problem of measurement error occurrence when direct current deviation occurs in the input and the amplitude is changed according to the original adaptive notch filter (ANF) method can be solved. Even though the amplitude and frequency of the input signal are changed suddenly and even the direct-current deviation is added, the fundamental frequency, the direct current deviation, and each harmonic instantaneous value and each harmonic amplitude can be extracted in real time with high precision based on the algorithm.

Embodiments of the present invention provide a road structure information extraction method, unmanned aerial vehicle, and automatic driving system. The method includes: obtaining at least one frame ofimage data of the road (S201); determining a semantic map of the road based on the at least one frame of image data (S202), and determining road structure information of the road according to the semantic map and the road structure model, wherein the road structure information includes lane grouping information (S203), which realizes the automatic extraction of road structure information withoutmanual annotation, and can extract road structure information in real time, which improves the extraction efficiency of road structure information and the road structure information includes not onlylane information but also lane grouping information, and description of the road structure is more specific. .

The invention discloses an unmanned aerial vehicle image ground fracture identification and extraction method based on machine learning. The method comprises the steps of obtaining mining area land crack image data through unmanned aerial vehicle photogrammetry, each image data is cut into small image data with equal pixels, constructing an image data set with different background information according to clustering analysis, so that a learning sample of a land crack recognition model based on machine learning is established, and verifying the classification accuracy of the model in a cross manner by adopting a machine learning algorithm of a support vector machine and a one-leaving method. When crack information is extracted, identified crack-free images are changed into image data with afull-white background, crack extraction is carried out on identified crack-containing images to obtain gray scale image data with a white background and black crack information, and the two types ofimages are spliced according to a cutting sequence. According to the method, the problem that the application of a machine learning algorithm is limited due to the fact that mining area land surface information is too complex in mining area ground fracture recognition application is solved, and the real-time performance, precision and efficiency are high.

The invention discloses a single-point noise resistant method for positioning the light spot center, which comprises the following steps of: comparing the gray value of each pixel of a light spot image to a preset threshold value; marking the pixel determined to be lower than the threshold value as a background pixel; for the pixel determined to be higher than the threshold value, judging if the adjacent pixel thereof is the background pixel; if yes, marking the present pixel as a noise pixel; if not, marking the present light spot pixel according to the mark of the pixel adjacent to the present pixel; according to the mark of the adjacent pixel, performing center-of-mass parameter set value accumulation and storage on the light spot pixel, and performing no center-of-mass parameter set value accumulation and storage on the real noise pixel; and when determining that each pixel is marked, calculating the center-of-mass coordinate of each light spot according to the final accumulated value of the center-of-mass parameter set of the light spot. The invention also discloses a single-point noise resistant device for positioning the light spot center. By using the method and device, the noise point is removed while the light spot center is extracted, the data storage space is saved, and the light spot center is extracted in real time.

The invention provides a method for early warning monitoring and analysis of enterprise complaints based on an artificial intelligence consumer right protection index, including the following steps: 1) using a data acquisition module to obtain the work order contents of the 12345 hotline; 2) establishing a classification module for the step 1), performing classification via an embedded classification module according to keywords through the work order contents acquired by the data acquisition module in the step 1) to obtain secondary indicators, and performing grouping and classification for important levels; 3) establishing the scoring rate for calculating the secondary indicators by an expert scoring module; 4) constructing an SVM classifier; 5) building a training classification model;6) building a classification carding module to perform service classification sorting on work order data; and 7) constructing a monitoring module for early warning monitoring and analysis of enterprise complaints combined with index and service data.

The invention discloses a range finding method and a range finding device based on a binocular video. The range finding method comprises the steps of acquiring a binocular image; performing preprocessing on the binocular image; finding a corresponding point which matches each pixel point from the binocular image through a stereo matching algorithm, calculating a parallax of each pixel point of the binocular image and obtaining a parallax graph; setting a sample set and initializing the sample set by means of a processed first parallax frame; performing parallax optimization and target extraction on the current parallax frame by means of the sample set, and updating the sample set, and utilizing the sample set in parallax optimization and target extraction on a next parallax frame; calculating the average value of parallaxes of all pixel points of a target extracted from the current parallax frame, and calculating for obtaining a distance to the target according to a binocular range finding algorithm. According to the range finding method and the range finding device, after the parallax graph is acquired through the stereo matching algorithm, algorithm precision is improved through processing of multiple frames, and furthermore the target is effectively extracted. The range finding method and the range finding device can settle a problem of incapability of simultaneously realizing high efficiency and high precision.

The invention provides a device and a method of multispectral imaging by diffraction based on a Nipkow disk, belonging to the technical field of imaging spectroscopy. The device comprises a diffraction lens, a chromatography and imaging system consisting of a light shield, the Nipkow disk and a micro-lens array disk, a detector, a relative position fixing mechanism, a first and a second rotation driving mechanisms and a linear actuator, wherein, when the relative position fixing mechanism is unlocked, the first and the second rotation driving mechanisms can drive the light shield, the micro-lens array disk and the Nipkow disk to rotate respectively; when the relative position fixing mechanism is locked up, a first driving mechanism can drive the chromatography and imaging system to scan the focal plane in the direction of an incident image so as to directly acquire the two-dimensional image and the spectral information of the incident ray at a single wavelength; and the linear actuator can drive the chromatography and imaging system to scan along the optical axis of the diffraction lens so as to acquire the image and the spectral information of incident rays at all wavelengths within the detected spectral range. The invention has the characteristics of low stray light, high spectral resolution and high spatial resolution.

The invention discloses a slit-lamp anterior segment sectional image feature extraction method comprising the following steps: 1, using a slit-lamp to gather an anterior segment sectional image so asto obtain an anterior segment color image; 2, carrying out color space transformation for the anterior segment color image, carrying out multi-threshold binary, distinguishing cornea, iris and crystalline lens latent areas, and respectively using blob shape analysis to obtain the cornea, iris and crystalline lens areas; 3, using morphological operation to carry out hole filing and expansion process for the cornea, iris and crystalline lens areas; 4, using a level set algorithm to track the boundaries, precisely tracking surface boundaries of each part of the anterior segment, and finding the anterior segment. The method solves the problems that the level set algorithm is slow in speed when applied in the whole image, can realize real time extraction and analysis of the anterior segment sectional image features, thus providing reliable basis data for solving anterior segment clinic parameters.

The invention discloses a multimedia digital watermark evidence taking mobile phone. The multimedia digital watermark evidence taking mobile phone is an intelligent evidence taking mobile phone whichuses the unique stability, concealment, security and fragility of a fragile digital watermark to ensure the integrity, the authenticity, the stability, the reliability and the authority of photographed and recorded multimedia data based on an image, audio and video fragile digital watermark embedding and extraction technology. When a user uses the mobile phone, imperceptible and unmodifiable fragile digital watermark information with unique identity confirmation is embedded in real time in a photographed image, audio and video. The image or the audio or the video with the fragile watermark canbe normally used and distributed but can not be modified manually. The authenticity of a file can be identified according to whether the photographed digital file contains the digital watermark information. The evidence taking process is quick, efficient, accurate, safe and convenient. The evidence taking content has the uniqueness of a digital evidence and the legal authority of evidence use, and the multimedia digital watermark evidence taking mobile phone disclosed in the invention can be used for evidence taking and recording in fields such as judicature, law enforcement and claim.

The invention brings forward a ZY-3 image-based shallow lake seine zone extraction method comprising the following steps: ZY-3 images of a research area are obtained and are subjected to image pre-processing operation; according to the pre-processed images, a typical seine sample database is chosen; the pre-processed images are subjected to gradient transformation operation; based on the transformed images, mean value and variance of seine sample zones are calculated; a seine zone threshold value is determined according to the mean value and the variance of the seine sample zones, and a seine zone is preliminarily extracted; a seine zone extracting result is optimized; optimization operation comprises operation of removing salt and pepper noise parts and low frequency parts and operation of removing broken spots; geometry characteristics of the seine zone are calculated, interpreted distribution and area of the seine zone are used as reference, and the extraction result is subjected to precision verify operation. Via use of the extraction method disclosed in the invention, a seine can be rapidly extracted in real time in a large scope and high precision manner, a data foundation is provided for further researching relations between shallow lake seine area and water environment, and basis is provided for making a water quality optimizing and lifting solution.

The embodiment of the invention discloses a passive visual system-based molten pool edge extraction method. A region and an edge in a molten pool image acquired by a passive visual system are modeled respectively, the two models are fused into a region-based active contour model and an edge-based active contour model respectively, the two active contour models are combined, and the combined active contour models are solved to accurately extract the edge of a molten pool in real time and provide useful information for the regulation of welding parameters in an automatic welding process.

The invention discloses a camera and three-dimensional laser radar data fusion-based road edge detection method. The method comprises the following steps of 1) establishing a grid map by taking a three-dimensional laser radar as an original point; 2) extracting lane line feature points through a camera by adopting an existing method, and measuring transverse and longitudinal distances of the feature points relative to an original point of the camera on an actual road; 3) acquiring road edge candidate points by using a method based on adjacent point height difference characteristics, 4) projecting the lane line feature points to a grid map according to the position relationship between the camera and the laser radar, and fitting by using a quadratic curve model, 5) judging the road edge candidate points within a fixed distance threshold from the lane line curve as road edge points, and 6) filtering the road edge points by using an RANSAC algorithm, and fitting the road edge points in combination with a quadratic curve model. The method can improve detection accuracy of the road edge, and thereby the accurate and reliable road boundary information is provided for the unmanned vehicle, and the occurrence rate of road traffic accidents is reduced.