148results about How to "Fast output" patented technology

The invention discloses a forest fire audio recognition method based on a BP neural network, a medium and a product in the technical field of forest fire recognition. The method comprises the following steps: acquiring forest fire audio data to be detected; extracting a Mel-frequency cepstrum coefficient (MFCC) based on the frequency domain features of the forest fire audio data to be detected; inputting the Mel-frequency cepstral coefficient MFCC into a pre-trained forest fire audio recognition model based on a BP neural network, and outputting a forest fire category; wherein the forest fire category comprises a forest fire and a non-forest fire. Compared with a traditional visual monitoring means, the method can cover a wider forest area, in the same monitoring range, in addition, model training based on the BP neural network can efficiently process a large amount of audio data and rapidly output an identification result, the efficiency and accuracy of forest fire identification are further improved, and the method is suitable for popularization and application. The problems of limited coverage range, high cost and low accuracy in the existing forest fire identification technology are solved.

The invention is suitable for the technical field of energy consumption analysis, and particularly relates to a smart energy consumption analysis method and system, and the method comprises the steps: drawing a technological process PID diagram, marking energy consumption equipment, positioning a control node of the energy consumption equipment, defining a standard working condition, reading a valve position of the control node through a preset DCS central control system, and carrying out the control of the energy consumption equipment. The fluctuation range of the valve position under the standard working condition is configured; and traversing the technological process PID diagram, determining the energy type of each control node, and clustering the fluctuation range based on the energy types. By generating the monitoring tree, the energy analysis result can be quickly output, the abnormal node can be found out, the energy consumption abnormal position can be accurately positioned, the more effective energy consumption management strategy can be formulated, and by drawing the change trend chart, the energy consumption condition can be visually displayed, the energy consumption state can be comprehensively known, and the energy consumption abnormity can be timely found and processed. And stable operation of a production system is ensured.

The invention belongs to the technical field of electric power engineering monitoring, and discloses a small sample power transmission line pin defect detection method based on a graph convolutional network, and the method comprises the steps: collecting a power transmission line pin image through image collection equipment; marking pins in the image based on event logic guidance to obtain a transmission line pin data set with logic association; introducing a cross-domain interaction enhanced pyramid network into the graph convolutional network to obtain an improved graph convolutional network model, and training to obtain an original target detection model; a knowledge graph relation reasoning module is introduced into the original target detection model, and a power transmission line pin defect detection model is obtained; and detecting the preprocessed power transmission line pin image by using a pin defect detection model, so as to generate a pin image with marked pin position classification and confidence. The method can effectively solve the problem that the sample size is small, and improves the recognition accuracy of the pin and the defect of the pin, thereby facilitating the improvement of the efficiency of detecting the defect of the pin of the power transmission line by inspection personnel, and guaranteeing the stable operation of a power system.

The present invention provides a 5-hydroxytryptophan production strain and a construction method and application thereof. The strain has a clear genetic background and is effective in producing 5-hydroxytryptophan. E.coli HTP11 Subtilis subtilis spores pureF bsu K316Q Mutated gene, upregulated strain E.coli HTP11 of guaB , guaA , gmk , ndk The transcription level of the gene was increased by heterologous introduction of the gapX Gene, missing aromatic amino acid transporter serB , tnaB , aroP Gene, up-regulated inner membrane efflux transporter yhjV The transcription level of the gene was increased and electroporated into the pTm-serB plasmid, which could stably provide 5-hydroxytryptophan, the direct precursor of 5-hydroxytryptophan, and had a high efficiency of 5-hydroxytryptophan output.

The embodiment of the invention discloses a decoration effect picture generation method and device based on generative artificial intelligence. A specific embodiment of the method comprises the steps of obtaining a to-be-transformed indoor image and an expected style prompt word; interference term elimination is carried out on the indoor image to be transformed to obtain an indoor structure image; performing line extraction on the indoor structure image to obtain an indoor line image; performing depth identification on the indoor structure image to obtain an indoor depth image; based on a pre-trained decoration image generation model, performing effect image generation on the indoor structure image according to the expected style cue word, the indoor line image and the indoor depth image to obtain an initial indoor effect image; and performing local redrawing on the initial indoor effect picture to obtain an indoor decoration effect picture. According to the embodiment, the efficiency of generating the decoration effect picture is improved, and waste of computing resources is avoided.

The invention discloses a lithium ion battery fault diagnosis method based on lithium plating detection, comprising the following steps: obtaining data such as battery thickness increment, battery temperature, ambient temperature, charge and the like through experiments; establishing a basic model for battery lithium plating layer detection, a thermal expansion compensation model and a lithium embedding expansion compensation model; adding pressure measuring points to repeat the experiment and establishing a gasification expansion compensation model; establishing a lithium ion battery lithium plating layer detection model; inputting the real-time measured battery temperature and ambient temperature, charge and battery thickness increment into each compensation model to obtain a corresponding thickness compensation value, subtracting the thickness compensation values ​​of thermal expansion, lithium embedding expansion and gasification expansion from the real-time measured battery thickness increment value to obtain a basic lithium plating layer thickness value approximately measured in real time; establishing a lithium ion battery lithium plating layer thickness detection estimator, detecting the lithium plating layer thickness online according to the real-time measured battery thickness, ambient temperature, battery temperature and charge and the like data, and performing lithium plating fault diagnosis.

The invention discloses a fabric dyeing equipment running machine parameter optimization recommendation method and device, equipment and a medium, and the method comprises the steps: obtaining the cloth type information of a plurality of production batches of an order system and the production information of a corresponding batch in a production execution system, carrying out the coding of the cloth type information and the production information, and obtaining cloth type feature data and process feature data, associating the seed distribution characteristic data with the process characteristic data to form process data items, calculating the seed distribution information of a target production batch by adopting a similarity algorithm when the target production batch information is detected, screening out preliminary process data items in a similarity threshold range, and storing the preliminary process data items in a database; and calculating the process characteristic data of the preliminary process data items by adopting a GBDT gradient decision tree to obtain recommended optimal line machine parameter data, so that the line machine parameter data of the dyeing equipment can be quickly output through system recommendation, and meanwhile, the batch quality problem caused by personal errors is reduced.

The invention belongs to the field of mine unmanned driving virtual scene generation, and discloses a mine unmanned driving virtual scene generation method based on a large model, which comprises the following steps: collecting network information and mine data, fusing the network data and a mine data set, forming a pre-training data set, and obtaining a sample Token; the sample Token is subjected to two stages of pre-training and weight cyclic fine tuning in sequence, feedback correction and cyclic tuning until a final model generated by the mine unmanned driving virtual scene is output; and inputting a to-be-trained unmanned driving algorithm and a required text to the final model, automatically generating a marked mine unmanned driving data set, scoring a data set generation result, feeding back to the weight cycle fine tuning stage, providing a scoring reference for the data set, and performing cycle tuning to obtain a mine virtual generation final model. According to the method, the problems of data shortage, high data labeling cost, insufficient training data diversity, insufficient virtual scene fidelity, high equipment cost and the like of the mine unmanned driving technology are solved.

The utility model relates to a pull-type straw smashing and packing machine. Comprising a rack, a picking and smashing device installed at the bottom of the rack and used for replacing the functions of a straw picking device and a straw crusher, a dust removal mechanism connected with the picking and smashing device and used for smashing straw and removing dust, and an auger conveying mechanism used for outputting straw in the dust removal mechanism to a double-bag-outlet packaging device. The traction and power input system is used for providing power; the picking and crushing device is used for synchronously picking and crushing straws collected from the ground, the dust removal mechanism is used for removing dust from the straws in the picking and crushing device and discharging the dust to the ground from the bottom of the rack, and the traction and power input system is used for being connected with a traction vehicle and inputting power of the traction vehicle; the utility model has the advantages that the equipment volume is greatly reduced, and the efficiency can be effectively improved.

The invention relates to the technical field of electrolyte preparation, and discloses a lithium ion battery electrolyte capable of ensuring low-temperature discharge characteristic, which comprises the following components in percentage by mass: 25% of propylene carbonate (PC), 35% of dimethyl carbonate (DMC), 20% of gamma-butyrolactone (GBL), 20% of ethyl acetate, 1.0 mol / L of lithium salt bis (trifluoromethanesulfonimide) (LiTFSI), 3% of additive 1, 3-propane sultone (PS) (mass fraction: 0.1% of sodium hydroxide (mass fraction: 0.1% of sodium hydroxide), and the balance of water. According to the present invention, the low temperature performance of the solution is improved, the unique solvent formula and the unique solvent ratio enable the electrolyte to have good fluidity in the low temperature environment, and the freezing point of the electrolyte is effectively reduced so as to ensure that the lithium ions can smoothly migrate between the electrodes at the low temperature, such that the lithium ion battery performance is improved, and the lithium ion battery performance is improved. The discharge performance of the battery under a low-temperature condition is obviously improved, and the application scene of the lithium ion battery is expanded.

The invention relates to a reinforcing steel bar acceptance system and method based on a BIM technology and an image recognition technology, and relates to the technical field of project acceptance, and the method comprises the steps: obtaining the acceptance time of reinforcing steel bar meshes of a plurality of regions; if the current time node is the acceptance time of the reinforcing mesh in a certain area, acquiring the reinforcing mesh image for analysis and identification; acquiring actual position coordinates of each reinforcing steel bar in the reinforcing steel bar mesh image through a BIM (Building Information Modeling) technology; according to the actual position coordinates of each steel bar, the distance between the adjacent vertical steel bars and the distance between the adjacent transverse steel bars in the target steel bar mesh are calculated, and acceptance check results about the positions, specifications, lengths and the distances of the steel bars are output. Data resources generated by checking and accepting the reinforcing steel bars can be reduced, and the checking and accepting result of the reinforcing steel bars can be rapidly output.

The utility model relates to the technical field of raw material mixing equipment, and discloses raw material mixing equipment for producing a wear-resistant shock pad, which comprises a bearing frame, a sealed sealing groove is arranged at the upper part of the bearing frame, a conical mixing barrel for mixing is arranged in the sealing groove, and an inner cavity of the conical mixing barrel is in a conical shape with a wide upper part and a narrow lower part. According to the utility model, the inner cavity is in the shape of the inverted cone, the first driving motor drives the mixing frame to rotate, raw materials move towards the periphery under the action of centrifugal force, the diameter of the conical mixing barrel is gradually reduced from top to bottom, a gradient type constraint space is formed, and the raw materials at the inner bottom of the conical mixing barrel are pushed upwards; circumferential flow driven by centrifugal force is coupled with reaction force of the wall surface, so that a spiral ascending flow state is formed, the mixing efficiency of the raw materials is improved, unmixed raw materials are not left at the bottom in the conical mixing barrel, and the mixing efficiency of the raw materials is improved.

The utility model belongs to automatic packaging equipment technical field especially is a kind of hot melt adhesive folding box machine. Including folding box machine frame, the inside of folding box machine frame is provided with synchronous feeding mechanism and the material taking assembly located in the both sides of synchronous feeding mechanism, the top of folding box machine frame is provided with the material rack conveying component located in the upper synchronous feeding mechanism. The utility model is through modular layered design and automation collaborative control, realizes the full-process efficient operation of paperboard conveying, glue spraying, stamping and forming, synchronous feeding mechanism and material taking assembly accurate cooperation, ensure material positioning stability, mould stamping mechanism and forming mechanism vertical layout save space, spray gun component realizes even glue spraying, left hinge mechanism, right hinge mechanism and side hinge mechanism timing linkage, guarantee folding accuracy, export conveying line completes finished product rapid output, overall structure is compact, stable operation, have automation efficiency and hot melt adhesive process advantage, applicable to high-precision packaging demand.

The present invention discloses a high-quality biodiesel combustion performance evaluation system, including an experimental database retrieval module, a fitting formula calculation module and a machine learning prediction module. The present invention can predict the combustion characteristics and emission characteristics of a mixed fuel after high-quality biodiesel is mixed with other fuels through the use of the three modules, and can calculate the total error through the error curve to compare the accuracy of the fitting formula calculation and the machine learning prediction to determine the actual application method. The present invention mainly solves the problem of evaluating the combustion performance when high-quality biodiesel is added and applied, so as to provide feedback to guide the biodiesel preparation process and regulate and optimize the biodiesel components. At the same time, this invention is not limited to fuel mixed with high-quality biodiesel, and the combustion performance of any fuel type involved in the system after mixing can be predicted.

The invention discloses a spatial orientation real-time analysis method based on eye movement, relates to the related technical field of flight training, and aims to overcome the defects of low analysis accuracy and poor adaptability in the prior art. The method comprises the following steps: acquiring pilot eye movement data through eye movement tracking equipment, acquiring head space orientation data through a flight helmet space sensor, and acquiring environment data containing instrument visual complexity and aircraft attitude from a virtual flight simulator; performing time synchronization on the three types of data; extracting features such as coordinates of a fixation point from the synchronous eye movement data; comparing the head orientation data with the reference data to generate deviation data; inputting the eye movement characteristics, the deviation data and the environment data into an evaluation model; the model dynamically adjusts parameters according to environment data, and calculates pilot space orientation capability indexes in combination with eye movement characteristics and deviation data. According to the method, accurate integration of multi-dimensional data can be realized, the analysis comprehensiveness and reliability are improved, a dynamic flight scene is adapted, and real-time accurate evaluation is realized.

The application relates to the technical field of cognitive function assessment, in particular to a cognitive assessment method and system based on eye movement test. The cognitive assessment method based on eye movement test comprises the following steps: a cognitive task covering multiple cognitive subdomains is measured on a subject in a preset order, eye movement data corresponding to each cognitive task is obtained by taking a preset interest area of each cognitive task as a spatial constraint; quality control and preprocessing are performed on the eye movement data, multi-dimensional eye movement features are extracted and standardized; each cognitive task score and overall cognitive ability quantitative score are calculated to obtain standardized individual cognitive parameters; a normal population cognitive parameter distribution is taken as a benchmark to construct a multi-cognitive subdomain parameter distribution model, a subject cognitive state is judged by combining a machine learning classification model, and a mild cognitive impairment assessment result is obtained. The cognitive assessment method based on eye movement test can comprehensively depict the cognitive state of an individual and realize multi-dimensional joint assessment.

The invention discloses a map real-time construction and rescue path dynamic planning system oriented to a fire dense smoke scene, and relates to the technical field of fire rescue. The system comprises a data acquisition and processing module, an environment modeling and path construction module and a post-processing module, based on millimeter wave radar, laser radar and motion sensor data, multi-modal information is obtained, synchronized and converted, a semantic segmentation model is utilized to analyze environment perception data to generate a feasible space, and laser radar point cloud and motion sensor data are fused to realize real-time positioning, map construction and rescue path planning in a fire scene. The path is updated in real time according to fire scene environment change, and a generated result is visually presented through a post-processing module and communicates with a command platform for real-time interaction. The system has the advantages of long-distance sensing of the millimeter wave radar and short-distance high precision of the laser radar, can realize quasi-real-time and high-precision fire scene environment modeling and dynamic path generation, and provides reliable path reference for rapid action decision in fire rescue.

The embodiment of the present application provides a multi-channel output converter and a control method thereof, wherein the multi-channel output converter includes: a power conversion circuit and an output circuit, wherein the power conversion circuit includes a transformer or an inductor, and the output circuit includes at least two output units, wherein each output unit includes an output capacitor and at least one switch tube and / or a diode, wherein the at least one switch tube and / or a diode is connected in series with the output capacitor, and each output capacitor is used to connect different loads, respectively, and the method includes: according to the load energy required by the output unit of the corresponding channel, the energy stored in the transformer or the inductor is distributed to the corresponding channel for rated voltage or variable voltage output. The technical solution connects the switch tube in series with the output capacitor at the output end of the converter to form a multi-channel output, and controls it by adopting a dynamic energy distribution method, so as to realize a fast response output of multiple different voltages or variable voltages, has a simple structure, and reduces system losses, etc.

This application discloses a log processing method, apparatus, device, readable storage medium, and vehicle. The log processing method is applied to a log processing system and includes: based on the memory identifier of shared memory corresponding to each application process, searching for the shared memory and a first storage identifier corresponding to each memory identifier, wherein the first storage identifier corresponds to a first target storage space among M storage spaces, the first target storage space is used to store a first target log and a timestamp of the first target log, and the first target log is the log with the smallest timestamp among at least one unprocessed log generated by the application process corresponding to the memory identifier; obtaining the timestamps of the first target log from the N first target storage spaces corresponding to the N first storage identifiers, obtaining N timestamps; and outputting the log corresponding to the smallest timestamp among the N timestamps. According to the embodiments of this application, the processing speed of the log processing system can be improved, and logs generated by different processes can be quickly input in chronological order.

The invention relates to the field of bean curd production, and discloses a double-enzyme method whole-bean bean curd production method and an integrated device, and the method comprises the following steps: S1, soybean peeling treatment; s2, soybeans are crushed, and the crushed soybeans are put into a production integrated device to be roughly ground for the first time; s3, performing wet-process accurate grinding, namely grinding the crushed soybean-water mixture again, and circularly grinding for 3-5 times; s4, emulsifying and homogenizing by using a homogenizer, and boiling the soybean milk at high temperature; s5, filling the cooked soybean milk into a specified mold or box; s6, adding a plurality of enzymes for marinating, and packaging; s7, performing sterilization by a pasteurization method, and then performing thermal forming; s8, cooling and low-temperature storage are carried out, the grinding effect can be better, different materials are added in a matched mode, shaping is carried out in a box, a pressing block is not needed to extrude yellow serofluid and wrapping cloth, then sewage generation is avoided, microbial erosion is reduced, the shelf life is prolonged, and the whole bean curd is high in fiber, high in tenacity, smooth, tender and better in taste.

The invention belongs to the technical field of analog signal collection, and discloses a flight control computer telemetering analog quantity collecting and editing circuit which comprises an input signal conditioning circuit, an analog switch circuit, a voltage following circuit, an ADC sampling circuit, an FPGA master control module circuit and a PCM sending circuit. And after passing through the two voltage following circuits, the signal is output to the two analog-to-digital converters for time-sharing ping-pong alternate sampling, and then the signal is output to the FPGA main control module circuit by adopting a serial interface, and is output in an RS485 signal form after being isolated by a digital isolator. The invention has the advantages of bipolar and wide input range, avoids the measurement error caused by signal amplitude scaling, reduces the measurement error caused by the output impedance of a pre-stage circuit to the ADC sampling hold circuit, improves the sampling precision and the sampling speed, and also reduces the I / O resource occupancy rate of an FPGA (Field Programmable Gate Array).

The invention discloses a disposable schistosome egg extraction filtering component, belonging to the technical field of schistosome egg extraction. The scheme is realized under the control of an intelligent schistosome egg extraction system, through a flexible stirring mechanism, a multifunctional stirring rod is moved up and down while rotating and stirring, so that a stirring medium is quickly broken up and stirred and decomposed under flushing water, and then a coarse screen overflow alarm mechanism and a fine screen filter mechanism are used to filter the medium layer by layer, so that a sample and a filtrate are quickly separated and output under the action of negative pressure adsorption, thereby improving the filtering speed and efficiency, and the overflowed liquid is introduced into an overflow container by means of a filter membrane frame and an overflow pipe, and the clogging of a fine screen filter and a coarse screen filter is determined by the floating detection degree of an overflow detection float and an overflow detection float sensor rod, so that a staff member is warned to clean in time, thereby enhancing the filtering effect, and the filtering component is a multi-layer structure, which can be combined and separated, so that it is more convenient to use and has a good use effect.

The invention discloses a direct current charging interface identification and positioning method, and belongs to the technical field of automatic charging of electric vehicles. The method comprises a coarse positioning stage and a fine positioning stage. In the coarse positioning stage, the initial position of the charging port is quickly determined through image preprocessing, contour screening and ellipse fitting. In the fine positioning stage, a multi-factor cluster matching algorithm is adopted, and firstly, a matching preselected area of each feature hole is determined according to a preset geometric cluster relationship among a plurality of feature holes in the charging port; secondly, dynamically calculating scaling and rotation coefficients of the template to adapt to different postures; and finally, constructing a matching degree evaluation function in combination with a central gravitation and edge gravitation principle, and performing matching calculation on an original image and a gradient image template so as to realize accurate positioning of the feature hole. The method solves the problem that the recognition precision and efficiency of the charging port are difficult to consider at the same time in a complex light field environment, and has the advantages of high positioning precision, high processing speed, high environmental adaptability and easy engineering deployment.

This invention relates to the field of tunnel engineering, and particularly to a tunnel shotcrete design method, storage medium, and electronic device that considers spatiotemporal effects. The method includes: calculating the final value of vertical deformation pressure and the radius of the plastic zone of the surrounding rock without support based on tunnel foundation parameters and tunnel shotcrete design parameters; calculating the deformation pressure as a function of time using a deformation pressure calculation formula, combining tunnel excavation speed parameters, the radius of the plastic zone of the surrounding rock, and preset shotcrete hardening characteristic parameters; inputting the deformation pressure-time function into a shotcrete load-support structure model, calculating the safety factor of the shotcrete at each time node based on the shotcrete hardening characteristic parameters, and iteratively adjusting the tunnel shotcrete design parameters according to whether the safety factor meets preset requirements, until outputting tunnel shotcrete design parameters that meet the preset requirements. This invention dynamically calculates deformation pressure considering spatiotemporal effects and iteratively adjusts parameters to ensure the load-bearing safety of the shotcrete throughout its entire life cycle and reduces construction risks.

The application discloses a positioning end PPP-RTK real-time precision product monitoring method and device, and belongs to the technical field of satellite navigation and positioning. The method comprises the following steps: acquiring a PPP-RTK positioning result and a post-precise coordinate reference value in real time, calculating an ENU direction positioning deviation, fusing three types of parameters of satellite geometric configuration, product precision and data timeliness to generate a reliability index C, and dynamically adjusting an alarm threshold value according to the index C; establishing a positioning precision database to support rapid query and analysis; formulating a multi-level alarm rule for positioning continuity and precision overrun, and presetting a diagnosis item mapping relationship; automatically executing data recording, root cause diagnosis and information output after triggering the alarm, and realizing full-process automation from abnormal sensing to root cause diagnosis. Through the three-layer intelligent closed-loop system of "evaluation-prediction-diagnosis", the reliability, stability and operation and maintenance efficiency of the PPP-RTK positioning service are significantly improved.

The invention discloses an image data rotation optimization method and device, and the method is applied to a data rotation optimization network, and the method comprises the steps: obtaining packaged image data through an application layer, processing the image data into initial color space data containing brightness data and chrominance data, converting the initial color space data into a pointer type through a bridging layer; the code layer performs block matrix transposition operation on the brightness data and performs block matrix transposition and chrominance calibration processing on the chrominance data at the same time, and rotated brightness data and rotated chrominance data are obtained respectively; and combining the rotated brightness data and the rotated chrominance data into processed color space data in a code layer, returning the processed color space data to an application layer through a bridging layer, and outputting rotated image data. According to the invention, the overall processing duration of image rotation is reduced by adopting the block matrix while the matching between the bright and dark contours of the image and the color is ensured.

The invention relates to a video cover generation method and device, electronic equipment and a computer readable storage medium. The method comprises the following steps: reading a current video frame from a video file; performing face image detection on the current video frame, and setting a face image as a figure subject picture of a cover; recognizing face key points of the character main body picture, calculating the character main body picture of which the side face angle is smaller than an angle threshold value according to the face key points, and setting the character main body picture as a cover main body picture; identifying a text of the cover main body picture and a rectangular surrounding frame of the text, and setting the cover main body picture, in which the rectangular surrounding frame is not overlapped with the character main body picture, as a video cover picture; outputting the video cover picture as a cover picture of the video file; according to the technical scheme, intelligent judgment of the high-quality video frame can be realized, so that the available video cover picture is quickly output, and the efficiency of generating the video cover picture is greatly improved.

The invention relates to the technical field of facial mask preparation, in particular to a full-automatic high-speed facial mask machine which solves the problems that in the existing facial mask preparation process, automatic filling and packaging operation on facial masks are inconvenient, the packaging quality is affected by exhausting and sealing of packaging bags, stability of the packaging forming quality is inconvenient to keep, and the production efficiency is high. And the production efficiency is limited. A full-automatic high-speed mask machine comprises a machine frame body, an operation panel is installed at one end corner of the upper end of the machine frame body, a plurality of bag falling air cylinders are fixedly installed on the inner side of the upper end of the machine frame body, and four bag falling suction cups are fixedly installed at the output end of each bag falling air cylinder; a bag falling guide sleeve plate is installed at one ends of the four bag falling suction cups, and the bottom end of the bag falling guide sleeve plate is connected with an upper guide groove plate in a sliding mode. By filling and packaging the facial mask, the preparation efficiency of the facial mask can be effectively improved, and the filling amount and packaging can be monitored in real time, so that the stability of the preparation quality is kept.

The present application relates to a tea bag sorting method based on an improved YOLO model and a parallel robot, which integrates an improved lightweight YOLOv11 model and an optimal suction decision algorithm, cooperates with a double-robot execution end, realizes the full-chain intelligent upgrading of tea bag sorting from "perception" to "decision" to "execution", uses the improved deep learning model, greatly reduces the calculation amount while maintaining high precision, ensures the real-time matching with the high-speed production line, adopts the cooperative work of the double robot based on the optimal suction decision, makes the system throughput increase by more than 70% compared with the single robot system, not only can quickly output the tea bag position, but also innovatively provides the "confidence" quantitative index and direction judgment, realizes the leap from simple "detection" to fine "quality evaluation and classification", and has high economic value.

This invention provides a lithium carbonate price prediction method based on multimodal data collaborative driving, comprising the following steps: multimodal data source identification and data acquisition; multimodal data classification and standardization; personalized noise filtering and outlier correction of multimodal data; extraction of trend and correlation features from structured data; extraction of semantic and visual / speech features from unstructured data; extraction of key information and feature quantification from semi-structured data; weighted collaborative fusion and feature optimization of multimodal features; construction and training of a two-branch collaborative prediction model; model prediction and preliminary verification of prediction results; optimization and dynamic correction of prediction results; interpretability analysis and visualization of prediction results; and method effectiveness verification and continuous improvement. This method, based on multimodal data collaborative driving, addresses the pain points of existing lithium carbonate price prediction technologies, forming a comprehensive, high-precision, and highly practical prediction solution.