38results about How to "Solve difficult to extract" patented technology

The invention relates to a molten pool contour image extraction method for realizing a closed connected domain. The molten pool contour image extraction method comprises the following steps: 1, extracting a molten pool image; 2, preprocessing and calculating the molten pool image; 3, calculating based on a gradient graph convenient for matching a guide operator template; 4, performing image merging; and 5, extracting a contour line of the molten pool image. By utilizing the device and the method, the problem that the contour line of the molten pool is difficult to extract as the front end of the molten pool is influenced by arc light due to non-uniform brightness distribution of the front end and the rear end of the molten pool is solved; the problems of unobvious gray scale change and weak gradient of two end parts of the rear end of the molten pool after an image is converted into a grey-scale map due to dark area brightness and non-uniform distribution of the rear end of the moltenpool are solved; the indication effect is achieved on the welding seam forming and the welding quality improvement.

The invention relates to a CRNN-based Myanmao image text recognition method, and belongs to the field of natural language processing. The method comprises the following steps: constructing a trainingset, a test set and evaluation set data of Myanmao text information images; marking text information in the Myanmao image by using a Myanmao Unicode sorting algorithm; adopting a deep convolutional neural network to extract a corresponding feature vector sequence from the input Myanhuang language image; identifying the feature vector sequence obtained in the last step by utilizing BiLSTM in a recurrent neural network RNN, and obtaining context information of the sequence, thereby obtaining probability distribution of each column of features; and calculating all label sequence probabilities byutilizing CTC, and selecting a label sequence corresponding to the maximum label sequence probability as a final prediction result of the Myanmao language of each frame in the image based on the dictionary and the mode of searching the candidate target. According to the method, the recognition of the Myanmai image text is realized, the recognition accuracy is high, and the effect is good.

The invention relates to a method for extracting residual abamectin from crystallization mother liquor, which comprises the steps of removing ester impurities and further separating each component by using a C18 column or a silica gel column. The separation and extraction process is simple; the selected and used solvent is cheap, safe and environment-friendly, and meets the industrialized requirement; and the effective ingredient B1 has high yield and reaches about 90 percent.

The invention belongs to the technical field of genetic engineering and metabolic engineering, and discloses an escherichia coli genetically engineered bacterium for fermentation production of L-histidine, a method and application of the escherichia coli genetically engineered bacterium for fermentation production of L-histidine. The escherichia coli genetically engineered bacterium is prepared by taking wild escherichia coli MG1655 as a starting strain, and knocking out a gene lacI and a gene purR on a genome; integrating the gene prs, the gene zwf and the gene gnd; the method comprises the following steps: integrating genes hisG, hisD, hisB, hisC, hisA, hisF / H and hisIE; and a gene cgl2458 is integrated. According to the invention, a molecular biological technology is utilized to integrate histidine operon in a serratia marcescens mutant strain with high histidine yield into an escherichia coli genome, and an escherichia coli genetically engineered strain with high histidine yield is obtained by enhancing precursor PRPP supply, histidine efflux and other strategies, so that the problems of thick fermentation liquor, high histidine yield and high histidine yield can be solved. And the method has a relatively good industrial application potential.

The invention discloses a crack detection method based on a remote sensing image. The method comprises the following steps: firstly, carrying out gray processing on a high-resolution unmanned aerial vehicle aerial image; performing feature highlighting on the gray level image by using a black cap algorithm, performing thresholding on a black cap operation result, performing expansion on the binaryimage, performing corrosion operation on the binary image, and finally performing machine learning training after extracting a connected domain of the binary image so as to complete crack detection.The crack detection method is suitable for crack detection operation of the unmanned aerial vehicle remote sensing image under the complex background, a large number of experiments prove that the crack information of the unmanned aerial vehicle aerial remote sensing image can be rapidly detected, calculation is easy, and the crack detection method can be suitable for complex scenes.

The invention discloses a corneal reflection and optical coding-based human face living body detecting method and system, belongs to the security technical field of resisting human face recognition attacks and particularly relates to the processing and the analysis of corneal images. The method comprises the steps of representing each piece of binary password information by using brightness or darkness of light through an optical coding mode and reflecting the brightness and darkness change by a cornea; and obtaining corneal reflection information by extracting the brightness and darkness information of a corneal image and realizing the human face living body detection through information comparison. The method disclosed by the invention has the advantages that only the brightness and darkness information of the corneal image needs to be extracted and the problem that the corneal reflection information is difficult to extract can be solved; the method has very low requirements on camera pixels and later image processing and analysis, so that the method is easy to realize, economical and effective; and the method has very high security and common attacks of face recognition can be resisted.

The present invention relates to the preparation process of extract containing hypericin and flavone compounds. The technological process includes extraction with polar organic solvent, biphase extraction and acid precipitation and through the simple process the extract containing total effective component not less than 50%, hypericin of 2-4 % and general flavone up to 50-90 % may be obtained. Inthe technological process, composite stabilizer is used, so that the extracted product may be maintained stably.

The invention relates to the field of traffic planning, in particular to a traffic flow prediction method based on a dynamic graph. The method comprises the steps of setting a plurality of monitoring sensors for key nodes of a traffic road, monitoring and collecting traffic flow, road occupancy and speed data of the road for a long time, constructing a traffic flow prediction model based on a dynamic graph neural network, inputting historical traffic flow data, setting related hyper-parameters in the model, and predicting the traffic flow. The method comprises the steps of preprocessing input data, designing a dynamic graph updating algorithm, applying the algorithm to a dynamic graph neural network module, extracting spatio-temporal features by adopting the dynamic graph neural network module and a ConvLSTM module, outputting a prediction result after the features are fused, and finally training a whole model for traffic flow prediction. According to the method, the fitting degree of the model prediction output and the actual flow condition is better, the prediction output stability is better, no large fluctuation occurs, and the traffic flow prediction result is more reliable and has more advantages.

The invention relates to the technical field of legal text generation, in particular to an intelligent case generation method based on a judicial knowledge domain atlas, which comprises the following steps: S1, constructing the judicial knowledge domain atlas; s2, extracting case facts and judgment reason parts of the document; s3, classifying and splitting case facts of the document; s4, carrying out deep learning to generate an abstract; s5, correlation comparison of the content is carried out, and integrity and accuracy are verified; s6, optimizing a judgment key through a convolutional neural network model; s7, automatically generating correlation indexes with other historical cases according to correlation analysis in the judicial knowledge domain graph; s8, summarizing the results, and splicing the results into a complete case document; according to the method, by combining the judicial knowledge domain graph and the deep learning technology, the legal document is automatically generated, the integrity of document content and the accuracy of legal logic are ensured, historical cases are efficiently associated, and the efficiency and quality of judgment document generation are remarkably improved.

The invention discloses a multi-stage recovery treatment method and device for waste shiitake mushroom sticks. The method comprises the following steps: sterilizing, cleaning, drying and crushing thewaste shiitake mushroom sticks; performing fermentation and centrifugation to obtain a primary residual material and a concentrated solution; extracting the primary residual material, and then performing filtration through an ultrafiltration membrane to obtain a treatment solution; carrying out alcohol precipitation on the concentrated solution and the treatment solution, centrifuging to obtain aprecipitate, and carrying out spray drying to obtain lentinan. The device comprises sterilization equipment, cleaning / drying equipment, fermentation equipment, centrifugal equipment, extraction equipment, a reaction kettle, ultrafiltration equipment and spray drying equipment. The waste shiitake mushroom sticks are subjected to polysaccharide extraction in a mode combining alcohol extraction afterfermentation, the problem that the shiitake mushroom sticks are complex in component and difficult to extract is effectively solved, and the high and stable extraction rate can be effectively guaranteed.

The invention relates to the technical field of intelligent robots, in particular to an intelligent human-shaped accompanying robot and a voice interaction system thereof, and the system comprises a voice input collection and preprocessing module which collects a composite voice signal and a voice triggering instruction in a surrounding environment in real time through a distributed microphone array; the multi-sound-source voice separation module is used for combining and independently processing the potential sound source signals and separating the potential sound source signals into a plurality of independent sound source signals; the target user sound source identification and extraction module is used for predicting the change of a target sound source in continuous time frames by utilizing a dynamic Bayesian network and extracting a target user sound source signal; the semantic analysis and output module is used for generating a natural and smooth voice response through a voice synthesis technology; according to the method, the problem that the target user sound source is difficult to extract in a noisy environment, multi-sound-source overlapping and voice switching scenes is solved, and the robustness and the adaptive capacity of the system in a complex voice environment are greatly improved.

The invention relates to the technical field of petroleum exploitation, in particular to an oil exploitation method based on a nano motor and application of the nano motor. The oil exploitation methodbased on the nano motor comprises the steps that step 1, the nano motor is taken, and the nano motor is mixed with deionized water to prepare nano motor suspension; step 2, a hydrogen peroxide solution is taken and mixed with the nano motor suspension and then injected into the stratum of an oil well; step 3, a catalytic metal layer evaporated on the surface of the nano motor catalyzes hydrogen peroxide to be decomposed to generate an oxygen concentration gradient to enable the nano motor to move; step 4, when the nano motor moves in the stratum, hydrophobic oleophilic groups modified on thesurface of the nano motor are combined with oil drops, and the nano motor with the oil drops moves in stratum liquid; and step 5, the nano motor with the oil drops is extracted from the oil well, andthen centrifugal separation is carried out to separate the oil drops from the nano motor and the oil drops are extracted. The oil exploitation method can effectively improve the recovery ratio of an oil field.

The invention discloses an infrared small target feature enhancement method based on local motion perception, and the method comprises the following steps: 1, loading an infrared sequence image, selecting continuous T frames of images as network input, and extracting image features through employing a backbone network; 2, extracting rough target inter-frame motion information by using a rough motion estimation module CME, and generating a forward optical flow and a backward optical flow; 3, aligning a multi-frame sequence by using an energy enhancement module EnE in combination with an optical flow and a deformable convolution, and enhancing target features by using a 3D convolution with a convolution kernel size of 1 * 1; and 4, outputting an enhanced infrared image by using a feature enhancement task head for guiding optical flow learning, fusing multi-scale features through a target segmentation task head, outputting a target segmentation result, and further guiding learning of a feature enhancement network. The method can effectively adapt to scenes with dark and weak targets and background motion, and outputs a high-quality enhanced image of the infrared multi-frame motion small target.

The invention relates to the technical field of extraction of natural peptide in plants and in particular discloses an extraction method of wheat germ peptide. The extraction method of the wheat germpeptide comprises the following technological steps: a, adding purified water into wheat germ meal and stewing; b, carrying out enzymolysis to obtain desugared germ meal; c, adding sodium citrate; meanwhile, adding neutral proteinase, cellulose proteinase and flavored proteinase; carrying out the enzymolysis for 2.0 to 3.0 h; d, carrying out high-temperature enzyme inactivation; e, separating andpurifying. The extraction method has the advantages that the price of the germ meal as a raw material is low; an extraction process is simple and convenient and no organic solvent is added; the purityof active molecules of the obtained wheat germ peptide is high and the yield of the peptide is high, so that the extraction method has relatively high economic benefits.

The invention discloses a safety helmet wearing inspection method based on a YOLOv3 algorithm, and the method comprises the steps of carrying out the expansion of the detection scale of a YOLOv3 model, obtaining a feature graph with a larger size, and solving a problem that the detection precision is insufficient or the feature extraction is difficult when the features of a safety helmet are smaller. Meanwhile, a loss function in the YOLOv3 is improved, and the loss function is established based on an IoU value logarithmic loss between a scale detection frame and a real target frame, so that the problem that the loss function in a traditional YOLOv3 model is sensitive to scale information is avoided, and the efficiency and stability of iterative calculation of the YOLOv3 model are greatly improved; finally, the detection and extraction efficiency of the safety helmet features is higher, and the precision is higher.

The invention discloses a single-magnet-ring dual-processing-circuit switching wide-range leakage current sensor and method. A magnetic core works in a deep magnetic saturation state based on a magnetic modulation principle and is used for sensing a leakage current signal; the square wave self-oscillation circuit is used for sending a square wave signal to the coil by utilizing the inductance effect of the coil; each of the two signal processing circuits comprises a filter, a proportional amplifier and a lock-in amplifier, and only amplification factors of the proportional amplifiers are different, and the proportional amplifiers are used for processing induced leakage current signals; and the microprocessor is used for analyzing and processing the signal output by the signal processing circuit. The system can monitor the insulation state of the energy storage system in real time, effectively improves the safety and reliability of the system, and is suitable for photovoltaic power generation, wind power generation and other application scenes related to the energy storage system.

The invention relates to a detection method for low-frequency non-stationary signals. The detection method comprises the following steps: S100, determining a center frequency; S200, determining the bandwidth of a main lobe; S300, determining a sidelobe bandwidth; S400, adjusting a phase shift coefficient; S500, transferring a function; S600, determining amplitude / phase frequency characteristics; S700, if the phase requirement is met, entering the step S800, otherwise, returning to the step S400; and S800, ending. The method has the beneficial effects that the problems that inter-harmonics in a target frequency band are difficult to extract and interference is difficult to eliminate can be effectively solved, the 75-125Hz low-frequency inter-harmonics are well extracted through simulation and experimental verification of the detection algorithm, the extraction precision can exceed 50%, the extraction precision of second harmonics is up to more than 90%, the interference signal can be attenuated to below 1% (the amplitude can be suppressed to be close to 0), and the correctness and effectiveness of the detection algorithm are fully verified.

The invention discloses a target fluctuation characteristic characterization method based on coherent accumulation gain. The method comprises steps that step 1, radar receives echo data and records the echo data as X, the X is a J*K-dimensional complex matrix, and J and K are the number of distance units and the number of pulses of the radar echo data respectively; 2, partitioning of the radar echo data X on a pulse dimension is performed according to the length N, 3, a distance unit corresponding to a target to be analyzed on the lth pulse block is recorded as D1, and the rest steps. Comparedwith a method based on statistical distribution modeling in the prior art, the method is advantaged in that a problem that fluctuation characteristics of the weak target are difficult to extract is solved, the fluctuation characteristics of the weak target are represented, and thereby statistical analysis of the weak target can be conducted.

The invention discloses a P-wave extracting method based on orthogonal projection and independent component analysis. The P-wave extracting method is used for analyzing and processing electrocardiograph signals, and includes the steps that collected electrocardiograph signals are subjected to signal preprocessing; a QRS wave group template is obtained with the time domain average method; QRS wavegroup ingredients are filtered out with the space orthogonal projection method through the wave group template, and therefore the signal to noise ratio of P waves and T waves in residual ingredients is increased; the P waves and T waves are separated; P waves after separation serve as reference signals, and the P waves are extracted from original aliasing signals again with the semi-blind method.

The invention discloses a method and device for multi-stage recovery and treatment of discarded mushroom sticks; the method comprises: pretreatment of sterilizing, cleaning, drying and pulverizing waste mushroom sticks; fermenting and centrifuging to obtain primary residue and concentrated liquid ; The primary residual material is leached and filtered through an ultrafiltration membrane to obtain a treatment solution; the concentrated solution and the treatment solution are alcohol-analyzed and centrifuged to obtain a precipitate that is spray-dried to obtain lentinan; the device includes sterilization equipment cleaning / drying equipment Fermentation equipment, centrifuge equipment, extraction equipment, reaction kettle, ultrafiltration equipment, spray drying equipment; the present invention adopts the combined method of alcohol extraction after fermentation to extract polysaccharides from discarded mushroom sticks, effectively solving the problem of complex components of mushroom sticks. For difficult-to-extract problems, it can effectively guarantee a high and stable extraction rate.

The invention discloses a method for characterization of target fluctuation characteristics based on coherent cumulative gain, including the following steps: Step 1, the echo data received by the radar is denoted as X, where X is a J×K-dimensional complex matrix, and J and K are respectively The number of distance units and the number of pulses of the radar echo data; step 2, divide the radar echo data X into blocks according to the length N in the pulse dimension, and step 3, divide the corresponding distance of the target to be analyzed on the lth pulse block The unit is marked as D l , and the remaining steps. Compared with the existing method based on statistical distribution modeling, the method of the invention solves the problem that the fluctuation characteristics of the weak target are difficult to extract, and enables the fluctuation characteristics of the weak target to be characterized, so that the weak target can be statistically analyzed.

The invention belongs to the technical field of genetic engineering and metabolic engineering, and discloses a genetically engineered Escherichia coli strain for producing L-histidine by fermentation, a method and an application thereof, wherein the strain is a wild-type Escherichia coli MG1655 strain, and the genes lacI and purR are knocked out on the genome; the genes prs, zwf and gnd are integrated; the genes hisG, hisD, hisB, hisC, hisA, hisF / H and hisIE are integrated; and the gene cgl2458 is integrated. The present invention utilizes molecular biological techniques to integrate the histidine operon in a mutant strain of Serratia marcescens that produces high histidine into the Escherichia coli genome, and by strengthening the supply of precursor PRPP and strategies such as histidine efflux, a genetically engineered Escherichia coli strain capable of producing high histidine is obtained, which can solve the problem that the fermentation liquid is viscous and the product is difficult to extract, and has good industrial application potential.

The invention discloses a P-wave extracting method based on orthogonal projection and independent component analysis. The P-wave extracting method is used for analyzing and processing electrocardiograph signals, and includes the steps that collected electrocardiograph signals are subjected to signal preprocessing; a QRS wave group template is obtained with the time domain average method; QRS wavegroup ingredients are filtered out with the space orthogonal projection method through the wave group template, and therefore the signal to noise ratio of P waves and T waves in residual ingredients is increased; the P waves and T waves are separated; P waves after separation serve as reference signals, and the P waves are extracted from original aliasing signals again with the semi-blind method.

The invention discloses a face detection method and system based on corneal reflection and optical encoding, which belongs to the field of security technology against face recognition attacks, and specifically relates to the processing and analysis of corneal images. Through optical coding, each bit of binary code information is represented by light brightness or darkness, and the cornea reflects the brightness and darkness changes; the cornea reflection information is obtained by extracting the brightness and darkness information of the cornea image, and then the human body is realized through information comparison. Face liveness detection. The method of the invention only needs to extract the bright and dark information of the corneal image, which can solve the problem that the corneal reflection information is difficult to extract, and has very low requirements for camera pixels and image processing and analysis in the later stage, and the method is easy to implement, and is economical and effective; Security against general attacks on face recognition.

The invention relates to a Burmese image text recognition method based on CRNN, and belongs to the field of natural language processing. The invention includes the steps of: constructing training set, test set and evaluation set data of Burmese text information images; using Burmese Unicode sorting algorithm to mark the text information in Burmese images; Extract the corresponding feature vector sequence; use BiLSTM in the cyclic neural network RNN ​​to identify the feature vector sequence obtained in the previous step, obtain the context information of the sequence, and obtain the probability distribution of each column feature; use CTC to calculate the probability of all label sequences, Based on the dictionary and the pattern of finding candidate targets, the tag sequence corresponding to the maximum tag sequence probability is selected as the final prediction result of Burmese for each frame in the image. The invention realizes the recognition of Burmese image text, and has high recognition accuracy and good effect.

The embodiment of the invention discloses an information query method. The method comprises the steps of obtaining a to-be-analyzed programming statement; wherein the to-be-analyzed programming statement is generated by adopting a target programming language; analyzing the to-be-analyzed programming statement to determine a first syntax structure tree; wherein each node in the first grammar structure tree represents a grammar structure in the grammar structure tree; determining a basic data structure based on a programming statement corresponding to the target programming language; based on the basic data structure, reconstructing the first grammar structure tree to obtain a second grammar structure tree; wherein the second grammar structure tree comprises a data structure in the first grammar structure tree; and receiving a statement query request, and performing statement query from the second syntax tree based on the query request. The embodiment of the invention further discloses an information query device and equipment and a computer readable storage medium.

The invention relates to a method for extracting residual abamectin from crystallization mother liquor, which comprises the steps of removing ester impurities and further separating each component by using a C18 column or a silica gel column. The separation and extraction process is simple; the selected and used solvent is cheap, safe and environment-friendly, and meets the industrialized requirement; and the effective ingredient B1 has high yield and reaches about 90 percent.

The invention discloses a document-level geological spatial relationship extraction method and system, and relates to the technical field of geological data mining, and the method comprises the steps: marking geological text data according to a defined geological entity type and a geological spatial relationship type; inputting the labeled geological text data into a RoBERTa pre-training language model to obtain semantic features; inputting the semantic features into a convolutional neural network to obtain structural features; fusing the semantic features and the structural features; constructing a heterogeneous graph, processing the heterogeneous graph by using a graph convolutional network, and extracting node features of the heterogeneous graph; converting the heterogeneous graph into an entity-level semantic hierarchy graph, and obtaining a geological entity node vector containing path reasoning information; and predicting the spatial relationship between the geological entities to obtain a structured spatial relationship triple. The method can effectively solve the problem that the long-distance dependency relationship and the cross-sentence relationship are difficult to extract.

The invention relates to the technical field of extraction of natural peptide in plants and in particular discloses an extraction method of wheat germ peptide. The extraction method of the wheat germpeptide comprises the following technological steps: a, adding purified water into wheat germ meal and stewing; b, carrying out enzymolysis to obtain desugared germ meal; c, adding sodium citrate; meanwhile, adding neutral proteinase, cellulose proteinase and flavored proteinase; carrying out the enzymolysis for 2.0 to 3.0 h; d, carrying out high-temperature enzyme inactivation; e, separating andpurifying. The extraction method has the advantages that the price of the germ meal as a raw material is low; an extraction process is simple and convenient and no organic solvent is added; the purityof active molecules of the obtained wheat germ peptide is high and the yield of the peptide is high, so that the extraction method has relatively high economic benefits.