55results about How to "Overcoming identification" patented technology

The invention relates to a fiber link fault recognition method, a device and a system. The method comprises the following steps: detecting power of one or a plurality of optical network units in real time; when detecting a fault in the fiber link corresponding to a certain optical network unit, extracting at least two upstream optical power values in the fiber link fault occurrence process; and according to the upstream optical power values in the fiber link fault occurrence process, recognizing the fault type of the fiber link corresponding to the optical network unit. The fiber link fault recognition method of the embodiment of the invention detects whether a fault exists in the fiber link corresponding to the optical network unit in real time by detecting the power of the optical network units in real time. When detecting a fault in the fiber link corresponding to the optical network unit, the upstream optical power values in the fiber link fault occurrence process are extracted, and the fault type of the fiber link is recognized in real time according to extracted power values.

The invention discloses a dynamic gesture identification method in an interactive system, comprising the following four steps of: acquiring an image, and obtaining a current image frame after denoising; obtaining a moving region of the image by calculating the difference of the current image frame and a reference image frame, and obtaining a binary image by a skin color detection method; carrying out gesture segmentation on the binary image by a gray level projection method to obtain a gesture region and a barycentric position; and obtaining a final identification result by a classification identification method according to the barycentric position and fingertip characteristics. The method can be used for overcoming the problems of high complexity, low real-time quality and low identification rate in traditional algorithms, has the advantages of simplicity in implementation, good real-time quality and high identification rate, and can be applied to the dynamic gesture identification very well. Moreover, the method can tolerate the translation of a definition gesture, the scale and the transformation of a rotation angle, and has a good robustness.

The invention provides a method for filtering an ineffective drift signal of global positioning system (GPS) equipment. The method comprises the following steps of: recognizing and filtering an ineffective drift signal point produced by the static GPS equipment at first when an indoor signal or an outdoor signal is different; and then correctly recognizing an ineffective drift point without deleting normal trace points of a satellite signal difference by mistake by comprehensively using the GPS signal weakness and a space analysis result. By the method, the effectiveness of the GPS signal canbe determined accurately on a software layer; and ineffective positioning points can be filtered out to the maximum extent on the basis of guaranteeing the completeness of the normal trace.

The invention discloses a method for diagnosing vibration faults of a rotary machine. According to the method for diagnosing the vibration faults of the rotary machine, vibration data in various states are collected, process amount data in various states are correspondingly acquired at the same time, and therefore the vibration data and the process amount data in various typical vibration fault states of the rotary machine can be obtained. Characteristic values are extracted from the vibration data and the process amount data respectively to form characteristic value matrices, the characteristic value matrices completely describe the process varying rules of various typical faults, namely a sample matrix database of the various typical faults. On the basis, according to the method for diagnosing the vibration faults of the rotary machine, the image recognition technology is applied to the field of fault diagnosis, the defects of manual fault recognition and single characteristic quantity fault recognition in the present field of fault diagnosis are overcome through matching calculation of the characteristic value matrices, and high-accuracy automatic fault diagnosis is achieved.

The invention provides a separation device for coal and waste rocks. The separation device comprises a vibrating coal feeder, a conveying belt, a queuing device, an image sensor, an image collector, a microprocessor, an upper computer, a separation controller, a high-pressure spray gun, a coal briquette channel and a waste rock channel. Information of the image sensor is processed in real time through the microprocessor and then sent to the upper computer, image information preprocessing, feature extraction, soft measurement model establishing and separation signal outputting are carried out, separation signals are sent to the separation controller, raw coal naturally falls into the coal briquette channel, and the waste rocks fall into the waste rock channel through the high-pressure spray gun. The invention further provides a separation method. The separation device and the separation method have the advantages that the defects of a low recognition rate, heavy environmental pollution and the like existing in an existing coal and waste rock separation method are effectively overcome, the recognition rate of the separation method is high, the environmental pollution can be effectively lightened, the separation device is small in size, easy to install and maintain and low in cost, separation precision and the automation degree of the raw coal are greatly improved, and manual waste rock picking is replaced.

The invention provides a speaker recognition method combining a Gaussian mixture model and a quantum neural network. The method provided by the invention comprises the following steps: at the training stage, framing input training voice signals, extracting characteristic parameters and generating characteristic parameter vectors; then using a K mean value method and an EM (expectation-maximization) algorithm to obtain the Gaussian mixture model parameters of the characteristic parameter vectors, and finally utilizing the Gaussian mixture model parameters of all the training voice signals to train the quantum neural network; and at the recognition stage, obtaining the Gaussian mixture model parameters of a recognized speaker, then inputting the model parameters into the trained neural network, and obtaining a recognition result. The speaker recognition method is applicable to recognition of the speaker under the condition of less sample data and unbalanced sample data, and simultaneously the capability of the quantum neural network which can carries out effective judgment on voice data with cross data and fuzzy boundary of the speaker is utilized, so that the correct recognition rate of a system can be improved.

The invention discloses a method and a system for locating a moving target in a mine based on binocular CCD vision. The system for implementing the method comprises surface equipment and an underground device. The surface equipment comprises a base station controller, a location server, an Ethernet switch and a monitoring terminal. The underground device comprises a mine intrinsically safe base station and a wireless radio frequency identification tag. The method for implementing the system comprises the following steps: (1) using an RSSI algorithm to estimate the location of a moving target; (2) carrying out matching feature extracting and training on the wireless radio frequency identification tag of the moving target; (3) carrying out feature matching on images after sampling through an ORB algorithm; (4) using a binocular CCD vision sensor to stereoscopically calibrate the left and right images of the moving target; (5) getting an internal and external parameter matrix of the binocular CCD vision sensor, and calibrating the world coordinates of the moving target; and (6) carrying out location correction on the world coordinate information of the moving target, and getting the accurate location information of the moving target in a mine. The problem on how to accurately locate a moving target in a mine NLOS environment is solved. The reliability and robustness of the system are improved.

The invention relates to an intravascular optical coherence tomography intravascular wall contour automatic identification method. The method comprises the following steps: (1) preprocessing an original image Ori; (2) identifying the inner contour of the blood vessel wall, wherein the method comprises the following steps: step 1, carrying out corrosion operation; 2, sorting the areas; 3, carryingout morphological direction and area detection; 4, judging whether a guide wire exists or not; 5, identifying and correcting the guide wire; and 6, reconstructing an incomplete contour.

The invention discloses a system and a method capable of realizing operation identification during the monitoring of a remote desktop protocol (RDP). The system comprises an RDP virtual server-side module and an RDP virtual client-side module, wherein a protocol data flow-dividing processing module is established between the RDP virtual server-side module and the RDP virtual client-side module; the protocol data flow-dividing processing module transmits clear-text data to an RDP protocol data flow formatting module; the RDP protocol data flow formatting module processes RDP client-side data and RDP server-side data respectively; a virtual screen operation identification module analyzes effective operation and a return result of the effective operation; and the operation identification is finished during the monitoring of the RDP. By the invention, the defect that the conventional RDP technology cannot realize monitoring and the operation identification is overcome; by the system and the method, on the premise of realizing the operation monitoring of the RDP, the operation can be identified; and the whole system is simple and easy to operate.

The invention provides an automatic identification and sorting device for products with identification codes. The structural form of a traditional identification and sorting device is broken through. A transporting device transports the products with the identification codes along an assembly line; an identification code detecting device can detect the identification codes on the products to distinguish the products with the different identification codes; corresponding visible identifications are generated on the products by a visible identification generating device; and sorting is carried out on the products with the different identification codes by a sorting mechanism, and therefore the effect of the automatic identification and analysis can be achieved on the products with the identification codes. Due to the fact that most usual identification codes are fluorescent codes, two-dimensional codes and other identification codes which are difficult to identify by human eyes; and by means of the visible identifications generated by the visible identification generating device, the assistance can be provided for personnel or consumers to identify, check and distinguish the products with the different identification codes. Therefore, efficient identification and sorting can be carried out on the products, the convenience is brought to people to identify, check and distinguish the products, and the practicability is strong.

The present invention discloses a distributed data interaction based biological identification method and a distributed data interaction based biological identification system. The method comprises the steps of: S1, acquiring user biological data by a first terminal, judging whether the user biological data are effectively registered in the first terminal locally, if so, finishing identification verification in the terminal locally, otherwise, going to a step S2; S2, broadcasting the user biological data to other terminals in the system; and S3, after the other terminals receive the broadcast data, judging whether the user biological data are effectively registered in the terminals locally, if so, finishing identification verification in the terminals locally, and transmitting the verification result to the first terminal. The system comprises a plurality of identification terminals and monitoring centers. By adopting the method, by virtue of a distributed calculating method, a large number of complex storage and calculation works are distributed to a plurality of identification terminals, and the problems low system response speed and high identification error rate caused by centralized storage and calculation are overcome. The method can be widely applied to various biological identification systems.

The invention discloses a computer vision-based large field pest situation monitoring sampling device and a sampling method. The sampling device includes a pest killing device, a transparent pest carrying board for carrying dead pests from the pest killing device, two cameras which are located at the upper side and the lower side of the transparent pest carrying board respectively and are used for acquiring images of the dead pests, and a terminal used for receiving the images from the cameras and recognizing the images. With the device of the invention adopted, the problem of the stacking of the bodies of the pests in an automatic sampling process, and the problem of precision drop of a recognition model which is caused by difficult recognition of the images of the abdomens of the bodies of the pests can be solved.

The invention provides a novel multi-angle SAR image fusion imaging algorithm being a multi-angle SAR image fusion method with local image index optimization. According to binary segmentation results of a layover zone and a non-layover zone of image pixels, paired crossed classification and fusion processing, like larger value taking, smaller value taking, and criterion interchanging, is carried out on a layover zone image and a non-layover zone image, thereby eliminating the adverse influence on SAR image fusion by layover and shadow phenomena effectively. Meanwhile, image quality evaluation is carried out on each image after fusion processing, a local image index of each image is calculated by using a sliding window by taking each pixel as center, and a final fusion processing result Iopt based on a local image index optimization criterion. Therefore, the image index of the fusion processing result Iopt can be guaranteed not only to be optimal globally but also to be optimal in a neighbor region by using any pixel as center.

The invention discloses a traffic sign detection and recognition method based on HDR technology. The method comprises: step (1), adopting seamless fusion technology to carry out processing, which is shown in the description, on N LDR images of different exposure values to generate a multi-exposure HDR image; step (2), carrying out joint color and shape detection, and obtaining three sign types bysegmentation; and step (3), utilizing an SURF (Speeded Up Robust Feature) operator to respectively extract features of to-be-recognized sign images and standard sign images of the types correspondingthereto, calculating Euclidean distances of feature vectors thereof, and using names of standard signs, which correspond to minimum Euclidean distances, as names of current to-be-recognized signs. Themethod can adapt to extreme weather such as dense fog and rainstorms, and overcome impacts of poor illumination conditions such as driving a vehicle at night; makes up for deficiency of missed detection and error detection caused by only relying on color detection, and achieves especially obvious detection effects for certain signs, such as ''no parking'' and ''traffic lights ahead'', with special colors; and is faster in a running speed, and is also higher in a recognition accuracy rate.

The invention discloses a defect detection method for an industrial ray weld joint image. The method comprises a step of preprocessing a plurality of weld joint image data, then performing Fourier transform to obtain an amplitude spectrum thereof, and taking a first quadrant of the amplitude spectrum as image feature data, a step of marking the image feature data and dividing the image feature data into a training set matrix and a testing set matrix to be inputted into a classifier of a support vector machine, and training and testing the classifier to obtain the accuracy of the classifier inweld joint defect identification, a step of shooting an actual weld joint, preprocessing a weld joint image and performing Fourier transform to obtain an amplitude spectrum, and taking the first quadrant of the amplitude spectrum as image feature data, and a step of inputting the image feature data into the classifier, allowing the classifier to identify a weld joint defect, and judging that the weld joint defect exists if an identification probability is larger than or equal to a classifier weld joint defect identification probability, otherwise judging that the weld joint is normal. According to the method, the detection accuracy of the weld joint defect is improved, the detection cost is reduced, and the method has high adaptability and is suitable for most weld joint image detection and analysis.

The invention discloses a method for recovering power supply after stripping started by direct-current ground protection, which includes the following steps: (1) failure detection: a direct-current ground protection device is connected with a power supply system, and is used for detecting the failure of the power supply system; the power supply system comprises two uplink grid contact lines, two downlink grid contact lines and a plurality of feeder cabinets, and any feeder cabinet is connected with the two uplink grid contact lines or the two downlink grid contact lines; all the feeder cabinets are connected with the direct-current ground protection device, and when failure is detected by the direct-current ground protection device, all the feeder cabinets are tripped; (2) response processing: according to whether the feeder cabinets are connected with the two uplink grid contact lines and the two downlink grid contact lines at the same time, processing is carried out separately according to two conditions. By improving the failure recognition theory of the power supply system, the method for recovering power supply after stripping started by direct-current ground protection and the like, the invention can achieve the advantages of high failure recognition efficiency and high power supply recovery speed.

The invention discloses a communication signal modulation mode identification method, which overcomes the problems of low identification rate and limited identification types under a low signal-to-noise ratio, and comprises the following steps of: 1) forming a set of known signals: (1) segmenting the known signals; (2) performing time-frequency conversion on the N sections of signal data; (3) performing gray scale transformation on the N three-dimensional time-frequency graphs; (4) marking the N two-dimensional time-frequency graphs; 2) training a convolutional neural network: (1) setting theconvolutional neural network; (2) training the set convolutional neural network; (3) adjusting parameters of the convolutional neural network; (4) generating a trained neural network; 3) forming a setby the received unknown signals; and 4) identifying a signal modulation mode, wherein the convolutional neural network is formed by training according to the set formed by the known signals, and theset formed by the unknown signals is input into the convolutional neural network for recognition of the modulation mode to obtain the modulation mode of the signals.

The invention discloses a high-spectral-imaging-based field pest monitoring and trapping device. The device comprises a pest killing device, a sampling device and a terminal machine; the sampling device is a hyper-spectrometer camera; the device also comprises a pest discharge device, a conveying device and a power device; the pest discharge device comprises a housing whose top part is provided with a pest inlet connected to the pest killing device and bottom is provided with a pest outlet; a hollow mandrel which extends into the housing; a pest discharge wheel which is rotationally installed on the hollow mandrel and whose periphery is provided with pest discharge holes which are communicated with the inner chamber of the pest discharge wheel, and the inner chamber of the pest discharge wheel is communicated with a negative device through the hollow mandrel; a absorption stop valve which is fixedly mounted on the hollow mandrel and can seal pest discharge holes that are going to discharge pest. The device can solve the problems of inaccurate counting and variety identification due to overlapped pest bodies; by means of the device, not only the number of pest can be counted, but also the varieties, sex and instar of the pest can be identified, which can predict and discriminate field pest situation more accurately.

The invention relates to an extra-high voltage DC power transmission line single-end distance measuring method using a modulus maximum value ratio, and belongs to the technical field of relay protection of a high voltage DC power transmission system. The method comprises the following steps: acquiring fault voltages u<+> and u<-> of a positive line and a negative line after a fault, performing decoupling transformation on the fault voltages under different transition resistance and fault distances to obtain u1 and u0, carrying wavelet decomposition on u1 and u0 in terms of four dimensions to obtain first wave head amplitude of a line-mode component and a zero-mode component, calculating a ratio of a first wave head amplitude maximum value of the line-mode component to the first wave head amplitude maximum value of the zero-mode component under the same dimension and taking the ratio as a nerve network input quantity, training a nerve network by use of an input sample set and an output sample set (a fault distance arranged in simulation), and obtaining a fault distance measuring result by use of a test sample. According to the invention, as it is shown through simulation that the extra-high voltage DC power transmission line single-end distance measuring method using the modulus maximum value ratio is quite high in precision.

The invention discloses a visual recognition system of a direct-reading water meter and a recognition method thereof. The recognition system sets the corresponding numerical value 0 on ten arc segments indicating 0 to 9 on the circumferential surface of each character wheel of the direct-reading water meter. The two-dimensional code identification of ~9, the identification method includes the following steps: 1) taking the image information of the direct-reading water meter dial; 2) removing the noise in the image by using median filtering and Gaussian filtering on the dial image; 3) passing the two-dimensional code The positioning information of the logo is used to calibrate the image angle; 4) the threshold value of the dial image is calculated, and binarization is performed, and the gray value greater than the threshold is set to the maximum gray value 1, and the gray value smaller than the threshold is set to 0; 5) Recognize the value of each character wheel in the dial image separately, and combine the recognized values ​​of each character wheel into the water meter reading. The invention obtains the water volume reading of the direct-reading water meter by identifying the two-dimensional code, effectively solves the problem of insufficient memory of the single-chip microcomputer, and reduces the cost of the water meter visual system.

The invention discloses a pedestrian detection method based on YOLO neural network. The method of the invention uses artificially divided pedestrian detection scene pictures to learn YOLO neural network so that the YOLO neural network can recognize pedestrians. This method realizes a pedestrian detection method based on YOLO neural network. Compared with the existing methods, it has better detection effect and performance.

The invention discloses an Android malicious software detection method based on an API calling sequence, and mainly solves the problems that a confused application programming interface API cannot beidentified, extracted features cannot be replaced by APIs, the types of the APIs are various, and the calculation amount is large in the prior art. The method comprises the following steps: (1) generating a training sample set and a test sample set; (2) generating a small file; (3) generating a calling sequence set of the application programming interface API; (4) replacing an application programming interface API with the package name; (5) generating a feature matrix; and (6) detecting malicious software. According to the method, the API in the API calling sequence of the application programming interface is replaced by the basic package name, the confusion package name and the user-defined package name, so that the API types can be greatly reduced, the calculation complexity can be greatly reduced, the malware identification and detection efficiency can be improved, and the method can adapt to the replacement of API versions.

The invention provides an intravascular optical coherence tomography intravascular wall contour automatic identification method. The method comprises the following steps: (1), processing an original image Ori; (2), identifying the inner contour of a blood vessel wall, including morphological direction and shape detection; performing morphological opening operation; excluding a non-pipe-wall connected domain; performing segmenting when the catheter is connected with the catheter wall; reconstructing an incomplete contour; and retrieving the mistakenly removed signal; and (3), performing coordinate transformation.

The invention provides a processing method for an image of paper money. The processing method is implemented on an automatic deposit and teller machine with preset programs. The processing method comprises the following steps that (1) a money detection module of the automatic deposit and teller machine acquires the image of the detected paper money, and corrects the detected paper money; (2) the money detection module judges whether the effective image of the detected paper money can be acquired or not; if yes, a first number image of the detected paper money is further extracted, recognition and authentic identification are conducted on the image, and the image is stored; if not, down-sampling processing is conducted on the whole image of the detected paper money, and the image is stored; (3) the image stored in the step (2) is uploaded to an upper computer of the automatic deposit and teller machine. According to the processing method for the image of the paper money, the complete image of the paper money in an abnormal state or the abnormal paper money or object appearing in the money detection process of the automatic deposit and teller machine can be stored in real time, the problem that due to the paper money in the abnormal state or the abnormal paper money or object, stuck money cannot be easily analyzed is effectively solved, and transaction disputes which can be caused by the problem are avoided.

The invention provides a method for rapidly identifying H. villosa chromosomes in a wheat background. According to the method, two oligonucleotide probes are firstly provided, the nucleotide sequences of the oligonucleotide probes are respectively shown as SEQ ID NO.1-2, the two probes are mixed and combine with an ND-FISH technology and can be used for identifying the H. villosa chromosomes in a wheat and haynaldia villosa hybridization generation, 7 H. villosa chromosomes are accurately identified while wheat chromosomes and the H. villosa chromosomes can be clearly distinguished, and wheat-haynaldia villosa translocated chromosomes can be identified. By adopting the method, the shortcomings of a GISH technology and an existing ND-FISH technology can be overcome, and the effects of the GISH technology and the existing ND-FISH technology can be also achieved. The method facilitates identification of the wheat-haynaldia villosa translocated chromosomes and has the very good prospect in the field of wheat breeding.

The invention relates to an automatic detection system for the packaging quality of venous blood taking needles. The technical problems that an existing manual detection for blood collection tube packages is low in efficiency, large in occupied detection field and low in manual detection percent of pass are solved. The system is provided with a conveying belt, a transferring device, a detection workbench and a single-axis vision detection mechanical arm, and a movable suction cup set is arranged above the conveying belt; the detection workbench is provided with a bench body and a push plate. The bench body is provided with a transparent detection window group which can be turned over. The single-axis vision detection manipulator can reciprocate and is provided with a detection camera. Thesystem can be widely applied to online intelligent detection of the blood taking needle package.

The invention discloses a three-dimensional object pose estimation method based on PVFH characteristics, and relates to a three-dimensional point cloud object pose estimation method. The invention provides an improved global feature descriptor PVFH. The descriptor is used for achieving three-dimensional point cloud object pose estimation. The method comprises the steps of in an offline stage, rendering from the CAD model to obtain point clouds of a plurality of visual angles, and extracting PVFH features of a part of the point clouds to establish a model library, segmenting a point cloud object from a scene in an online identification stage, then extracting PVFH features of the identified object, carrying out feature matching with a model library to obtain a roughly matched pose, and finally carrying out optimization by using an ICP algorithm to obtain an accurate pose. According to the method, the defect that a traditional global feature descriptor cannot identify a rotationally symmetrical or mirrored symmetrical object is overcome, and the robustness is high.

The invention discloses a preparation method of L-phenylalanine print hydrogel based on zwitter-ion monomer. The preparation method comprises the following specific steps: performing alkylation reaction between 1-vinyl imidazole and 1,3-propane sultone to prepare the zwitter-ion monomer VSPIM; using L-phenylalanine as a template molecule, the prepared VSPIM as a functional monomer, N,N-methylene bisacrylamide as a cross-linking agent and water as a solvent, triggering polymerization through redox reaction and obtaining hydrogel; washing the obtained hydrogel in NaCl solution, measuring a characteristic peak of eluant at 259nm position through an ultraviolet spectrophotometer, continuing an eluting process until the characteristic peak in an ultraviolet spectrum of the eluant at the 259nm position cannot be observed and then eluting NaCl on the surface of hydrogel with deionized water to obtain the print hydrogel. The preparation method disclosed by the invention utilizes zwitter-ion radicals to reduce a nonspecific adsorption effect of print hydrogel on protein.

The invention discloses a photograph information processing and scenic spot identification method, a client and a server. The photograph information processing method comprises steps that the photograph attribute information of a newly-added photograph of the client is acquired; whether the newly-added photograph is shot by the client is determined according to the photograph attribute information; if yes, the shooting time and the shooting place of the newly-added photograph are uploaded and are taken as bases for scenic spot identification, and the purpose of automatic scenic spot identification is realized. The method is advantaged in that problems of untimely identification and high cost existing in a traditional artificial scenic spot identification mode in the prior art are solved, the technical effect of timely scenic spot identification with low cost is realized, and the technical problems of untimely identification and high cost existing in the traditional artificial scenic spot identification mode in the prior art are solved.

The invention discloses a short video label determination method, system and device and a storage medium, and the method comprises the steps: obtaining the audio information of a first short video, and carrying out the video-audio analysis of the first short video, and obtaining a first audio label; acquiring key frame information of a first short video, and performing video content analysis on the first short video to obtain a first scene label, a first object label and a first character label; acquiring title information, video description information and subtitle information of the first short video, and performing video semantic analysis on the first short video to obtain a first semantic tag; and performing weight decision analysis according to the first audio tag, the first scene tag, the first object tag, the first character tag and the first semantic tag to generate a first short video tag. The short video tag generation efficiency is improved, and the accuracy, comprehensiveness and reliability of the short video tag are also improved. The method can be widely applied to the technical field of video processing.