137results about How to "Avoid misleading" patented technology

The invention discloses a liquid fertilizer application effect analysis method based on self-supervised learning. The method comprises the following steps: S1, generating initial feature data; s2, environment feature representation is generated based on a heterogeneous environment self-adaption module; s3, extracting short-term and long-term spatio-temporal characteristics of the liquid fertilizer application effect based on the dual spatio-temporal characteristic decoupling and dynamic fusion model, and generating comprehensive spatio-temporal characteristic representation; s4, generating fertilization characteristics in different situations by utilizing self-supervised learning, and eliminating environmental interference factors through a causal reasoning module; s5, generating a fertilization intention feature, and dynamically adjusting a fertilization strategy; s6, when the difference between the prediction result and the actual effect exceeds a threshold value, triggering a re-learning module to update model parameters; and S7, carrying out integrated optimization, realizing feature sharing and task coordination through multi-task collaborative learning, and carrying out automatic tuning by utilizing Bayesian optimization. The intelligent analysis of the liquid fertilizer application effect is realized through the self-supervised learning and multi-task collaborative optimization method.

The invention relates to the technical field of gene analysis bioinformatics, and particularly discloses a gene data analysis method based on a large language model.The method comprises the steps that the quality of gene data to be analyzed is analyzed, whether the gene data to be analyzed is available or not is judged, and a gene sequence segmentation scheme of the available gene data is determined; inputting the gene data after gene sequence segmentation into the large language model to obtain a gene data association feature data set and a gene data analysis result, determining a gene data analysis accuracy feature value, and judging whether the gene data analysis result conforms to expectation or not; the problems that quality and integrity problems of traditional biological data challenge the accuracy of analysis results and effective integration and analysis of multi-modal data still need to be improved are solved, the data processing efficiency is optimized in the whole process, repeated calculation, misjudgment and invalid results are reduced, the analysis efficiency is improved, and resource waste is reduced.

The invention provides a control method and system for an electric drive system of an electric vehicle. The method comprises the steps that the motor rotating speed and the motor torque of the electric vehicle are continuously obtained; if the current rotating speed of the motor is larger than or equal to a first preset threshold value, it is determined that the electric vehicle is in a first operation stage, and if the current rotating speed of the motor is smaller than a second preset threshold value, it is determined that the electric vehicle is in a second operation stage; in the first operation stage, whether anti-skid control is carried out on the electric automobile or not is judged according to the change of the rotating speed of a motor of the electric automobile; and in the second operation stage, whether anti-shake control is conducted on the electric automobile or not is judged according to the current motor torque, if yes, torque compensation is conducted on the electric automobile according to a motor rotating speed oscillogram of the electric automobile, and the stability of the electric automobile in the running process is improved.

The invention discloses a bridge point cloud extraction method based on evolution gradient search, and relates to the technical field of data processing, and the method comprises the steps: employing a clustering algorithm to carry out the denoising of a bridge surface point cloud; preliminarily extracting the de-noised bridge floor point cloud based on a multi-thread and minimum heap data structure, and meanwhile, setting a perpendicularity density constraint to further extract the bridge floor point cloud; a gradient-based optimization algorithm is combined with an evolutionary strategy to serve as a complementary algorithm to be incorporated into a unified framework, and the extracted bridge deck point cloud is optimized by alternately adopting a gradient updating method and an evolutionary updating method, so that the finally extracted bridge deck point cloud is obtained. According to the method, the processing precision of the bridge deck point cloud is improved, the main body structure of the bridge can be more effectively identified and separated from the complex point cloud data by introducing the evolution gradient search strategy, and meanwhile, the interference of noise and redundant information on a final result is remarkably reduced.

The invention discloses an LLM data acquisition processing method and system based on an automatic capture technology, and relates to the technical field of artificial intelligence, and the method comprises the steps: distributing a data acquisition task for each crawling node in a distributed node architecture; the crawling node accesses different data sources based on the data acquisition task to obtain acquired data; after the crawling node completes a round of data acquisition, task allocation is adjusted based on the acquisition progress and the acquisition efficiency; performing first data processing on the acquired data, and performing data quality evaluation, invalid information removal and text correction on the acquired data after the first data processing; according to the method and the device, the collected data which is from different data sources and is subjected to the first data processing are integrated to obtain the integrated data, and the integrated data is subjected to the second data processing to obtain the LLM data, so that the problem that the data of various data sources for LLM cannot be efficiently collected and processed with high quality in the prior art is solved.

The invention provides a geoscience specimen digital acquisition and management system, and relates to the technical field of geoscience specimen acquisition, and the system comprises a data acquisition module which collects specimen attribute fields, specimen photos and process metadata; the data quality evaluation module is used for performing bidirectional modulation on the basic weight based on the dependency relationship set of the fields to obtain an effective weight, and calculating an integrity score based on the effective weight; calculating an accuracy score, a specification degree score and a timeliness degree score; calculating to obtain a comprehensive quality score; the data optimization suggestion module is used for generating an optimization suggestion according to the score of each dimension and the score of the comprehensive quality; and the data storage management module is used for storing specimen attribute fields, specimen photos, process metadata, quality evaluation results and optimization suggestions. According to the method, the quality problem can be recognized in time, targeted optimization suggestions can be automatically generated, and intelligent priority recommendation is provided for field acquisition based on historical quality statistical data, so that the quality and efficiency of geoscience specimen data acquisition are improved.

The embodiment of the invention provides a gas station user information and local service linkage method and device and terminal equipment, and belongs to the technical field of data processing. The method comprises the steps of obtaining first consumption information corresponding to a historical user in a target gas station in a target area and second consumption information corresponding to the historical user in a target shop in the target area, and obtaining initial evaluation information corresponding to the target shop; performing false detection on the initial evaluation information to obtain target evaluation information corresponding to the target shop; determining a target consumption model corresponding to a historical user in the target area according to the first consumption information, the second consumption information and the target evaluation information; obtaining third consumption information corresponding to the target user in the target gas station and fourth consumption information corresponding to the related shop in the target area; and determining a target consumption route corresponding to the target user by using a target consumption model according to the third consumption information and the fourth consumption information.

The utility model belongs to the technical field of cranes, and discloses a crane loading device and a crane. According to the crane load device, a hoisting hook is connected with a supporting transverse plate, the two ends of the supporting transverse plate are arranged in opposite grooves of a first panel and a second panel in a sliding mode, a pressure sensor is arranged below the end point of the supporting transverse plate, an alarm is arranged on the first panel or the second panel, and the pressure sensor is electrically connected with the alarm; the supporting transverse plate is in direct contact with the pressure sensor, and the pressure sensor is electrically connected with the alarm, so that the load condition of the crane can be monitored and fed back in real time, and once a measured value exceeds a safe load range of the crane, timely warning can be provided to help an operator find potential problems of the device in time; the possibility of accident occurrence is reduced, the false alarm condition can be effectively avoided, the accuracy of load detection is improved, and misleading of operators is prevented.

The invention discloses a main distribution micro multi-level collaborative optimization regulation and control method and system adapting to non-smooth characteristics. The method comprises the following steps: constructing a converter equivalent model and a power grid topological relation according to power grid measurement data; obtaining a distributed energy three-phase reactive power piecewise function based on the converter equivalent model and a droop control principle, and performing approximate fitting by using a fitting function to obtain a droop control function; micro-grid decoupling is realized by using a node replication method and a micro-grid adjacent relation; constructing a multi-level collaborative optimization scheduling model by using a droop control function, micro-grid decoupling and an optimization objective function; and outputting a main and distribution micro-grid regulation and control strategy by using an interior point method according to the real-time data of the power grid and the multi-level collaborative optimization scheduling model in combination with a time sequence consistency strategy. The method has the advantages that it is ensured that the final regulation and control result is based on the characteristics of the distributed energy and the global characteristics of the distributed energy while the iterative calculation amount is reduced, and both pertinence and efficiency are considered.

The invention relates to the technical field of medical consultation, and particularly provides an automatic question answering method and system based on a medical document, a terminal and a storage medium, and the method comprises the steps: obtaining question text data, and calling target paragraph data matched with the question text data from a database, the database being used for storing paragraph data divided by the medical document; utilizing a plurality of language models to generate a plurality of candidate answers based on the target paragraph data; segmenting the candidate answers into logic sentence sets, and performing duplicate removal and fusion processing on the plurality of logic sentence sets to obtain a target set; and inputting the target set into a pre-training language model to obtain answer text data. The answering accuracy of the medical consultation is improved.

The invention discloses a short-term electricity price prediction method and system based on an MIC-TCN-Transform model, and the method comprises the steps: obtaining electricity price prediction data from different data sources, and integrating the electricity price prediction data into a unified format and structure; performing integrity, continuity, consistency and rationality check on the data, cleaning unqualified data, performing feature extraction and selection on a data set, performing normalization processing on all the data, and segmenting the data set into a training set, a verification set and a test set; the method comprises the following steps: defining a TCN model and a Transform model, and creating an MIC-TCN-Transform model; an optimizer and a loss function are configured, the MIC-TCN-Transform model is trained by using the training set, and the MIC-TCN-Transform model is evaluated by using the verification set; and the test set is inputted into the trained MIC-TCN-Transformer model, and a prediction result is obtained. Reasonable configuration and utilization of renewable energy sources can be effectively promoted, and the reliability and stability of a power system are improved.

The invention belongs to the technical field of power equipment state monitoring, and discloses a transformer discharge detection method and system with an anti-interference function, and the method comprises the steps: obtaining a multi-mode discharge original signal of a transformer, carrying out the synchronization of a coefficient through a channel, anchoring a synchronous interference signal, extracting initial multi-dimensional features, and carrying out the detection of the multi-mode discharge original signal. Performing weighted average fusion on the initial multi-dimensional features according to a channel synchronization coefficient to obtain an anti-interference feature vector; constructing a feature screening-interference immune lightweight neural network model, inputting anti-interference features, taking a channel synchronization coefficient as an interference immune factor, and outputting a discharge type identification result; and obtaining an evaluation index by combining the peak amplitude of the multi-mode discharge original signal of the transformer after the de-noising preprocessing, and obtaining a discharge risk level based on the evaluation index. According to the invention, accurate identification and risk grade evaluation of transformer discharge are realized, and the problems of poor anti-interference performance and the like are solved.

The invention relates to the technical field of automatic detection of automobile plastic injection molded parts, and specifically discloses an automatic detection system for automobile plastic injection molded parts based on dual robot technology, the system comprising: an injection molded part placement position assessment module, an injection molded part image quality assessment module, an injection molded part injection molding quality analysis module, a database and an injection molded part quality sorting module; the invention conducts a refined assessment on whether the placement position of the automobile plastic injection molded parts is accurate, and adjusts the placement position, and at the same time, flexibly collects images of various parts through a track-type detector, conducts an in-depth analysis of the quality of the collected images, and finally considers that the importance of the injection molding quality of each part in the comprehensive quality assessment is different, thereby reducing the possibility that the injection molded parts with inaccurate positions may cause the production line to stagnate, avoiding treating the defects of the secondary parts as the same as the defects of the main parts, thereby avoiding misleading the judgment of the overall quality.

The invention discloses a multi-modal tampering information detection and positioning method based on mitigation of modal misleading. The method comprises the steps of S1, image tampering region identification, S2, non-tampering region and text interactive learning, S3, LLM knowledge supplementation, and S4, tampering identification and positioning. After a tampered area and a non-tampered area of an image are recognized, an LLM generates explanation of text modal background knowledge, and the consistency of the text modal and the background knowledge and the consistency of the image modal and the background knowledge are compared to serve as the basis of multi-modal tampering information detection and positioning, so that the modal misleading of multi-modal tampering information detection and positioning is relieved.

The application belongs to the technical field of mechanical fault diagnosis, and discloses a rolling bearing small sample fault diagnosis method and system based on an online soft label Gaussian prototype network. The method collects bearing vibration signals under different working conditions to construct a data set, divides the data set into a meta-training set and a meta-testing set, and splits the data set into a support set and a query set; a Gaussian prototype network model containing embedding, prototype calculation and classification modules is constructed; in the meta-training stage, soft labels are dynamically generated by using an online soft label strategy, hard label loss is combined to carry out multi-task training to obtain optimal parameters; in the meta-testing stage, Gaussian prototypes are constructed based on the support set, classification is realized by calculating the Euclidean distance between samples and prototypes, and when the feature distribution drifts, the prototype parameters are fine-tuned to adapt to cross-condition diagnosis. The method effectively alleviates label noise interference, can still guarantee diagnosis accuracy under a small number of labeled samples, and significantly improves the model generalization ability and diagnosis stability.

The invention discloses a silicon controlled rectifier forward and reverse bidirectional switching solid-state module, and aims to solve the problems that an existing silicon controlled rectifier switching scheme is high in short circuit risk, high in cost and complexity and difficult to maintain, and a simple relay replacement scheme is short in service life. According to the technical scheme, the device comprises a hardware circuit structure and a single-chip microcomputer control logic part, the hardware circuit structure comprises a main loop silicon controlled rectifier unit, a relay switching unit, a single-chip microcomputer control unit, a signal acquisition unit and a power supply unit, and a mixed topological structure of three main loop silicon controlled rectifiers and a common double-pole double-throw relay is adopted. And in cooperation with closed-loop control logic of load state judgment, silicon controlled rectifier turn-off, relay switching and silicon controlled rectifier conduction, current forward and reverse safe switching is realized. Two silicon controlled rectifiers for switching are omitted, so that the risk of short circuit is thoroughly avoided; the relay is switched only when the silicon controlled rectifier has no output, so that the service life is prolonged; the circuit structure is simplified, the hardware and maintenance cost is reduced, and the anti-interference capability is remarkably improved.

This application discloses a GIS high-voltage switch housing with a safety protection structure for use in the field of high-voltage switches. This switch housing, through the inclusion of a sealing protection mechanism, allows the mechanism to monitor the sealing performance of the main housing. If a problem occurs in the sealing performance of the main housing, leading to SF6 failure... 6 When a gas leak occurs, the trigger board will activate the leak detection switch, causing the alert control module to activate the alarm and alert the staff. This allows staff to detect the gas leak in a timely manner and prompts them to conduct appropriate inspections and maintenance. This provides a protective effect, reducing the impact of problems with the casing's sealing. Furthermore, through the installation of gas sensors and gas analysis modules, when the leak detection switch is triggered, the sealing protection mechanism will also perform a self-check on its internal sealing to prevent false alarms. This avoids misleading staff and greatly improves practicality.

The utility model provides an anti-counterfeiting packaging film which comprises an AL film, an OPP film, a VC film, a false anti-counterfeiting mark, a covering anti-counterfeiting piece and an anti-counterfeiting two-dimensional code, the AL film, the OPP film and the VC film are compounded together, and the false anti-counterfeiting mark is printed on the OPP film; an anti-counterfeiting two-dimensional code is arranged on the AL film, and the AL film is compounded with the covering anti-counterfeiting sheet; according to the utility model, the AL film, the OPP film and the VC film are compounded, so that the composite film has protection and mechanical properties; the anti-counterfeiting mark and the anti-counterfeiting two-dimensional code are arranged in a staggered manner, and dynamic anti-counterfeiting of the thermosensitive covering anti-counterfeiting piece is matched, so that the anti-counterfeiting reliability and concealment are improved, and the anti-counterfeiting device is not easy to counterfeit.

The application discloses a power distribution box remote inspection management method and system based on an Internet of Things, and relates to the technical field of power distribution equipment operation and maintenance management. In view of the problem that remote inspection defaults to adopt sensing data, and after sensing link failure, back transmission of smooth false data forms virtual inspection, the method carries verification incentive configuration in the inspection instruction, executes switching of incentive objects selected from existing controllable loads in the power distribution box by the intelligent gateway, and synchronously collects the output of each sensor in the verification observation window; item-by-item comparison is performed according to three types of expected response modes of change, slow change and invariance of the incentive response verification table, link credibility and link abnormality are distinguished; only when verification is passed, the inspection data is adopted, an inspection work order is dispatched for an abnormal channel, and previous data is marked for rechecking according to a tracing period, and if the abnormality exceeds an upper limit, manual inspection is performed, remote alarm is temporarily stopped, and the inspection interval is adaptively adjusted according to the verification passing condition. The application can verify the authenticity of the sensing link without adding hardware, and is suitable for remote inspection of the power distribution box.

This disclosure proposes a method for synthesizing multimodal data for decision-making attribution explanation in the mining industry, comprising the following steps: First, extracting causal logic from industry texts to construct a structured rule base; second, collecting on-site multimodal data and filtering out data pairs strongly correlated with rules through scenario-based weight modeling and cross-modal association mechanisms; third, using a multimodal generative adversarial network with rule constraint mechanisms and combined with rule mutation technology to synthesize multimodal data with complete logical chains and covering various complex working conditions; finally, ensuring data quality and achieving self-evolution of the knowledge base through a three-level verification and rule feedback mechanism, and storing qualified data in a hybrid storage database. This invention solves the problems of complex causal relationships and scarce high-quality training samples in the mining field, and can automatically and in batches generate logically rigorous and scenario-diverse multimodal data, improving the training efficiency and generalization ability of attribution decision models.

The invention discloses a disordered stacking scene long-sequence object grabbing and releasing decision method based on a double-loop strategy, and the method comprises two closed-loop strategies, the inner-loop strategy is responsible for improving the object matching accuracy through active sensing, after a mechanical arm grabs an object, the strategy can sense the object from multiple visual angles through rotating the object, and the object matching accuracy is improved. And the matching confidence of the object and the object in the target scene is increased. The outer ring strategy is responsible for executing grabbing and placing operations, predicting grabbing actions in combination with uncertain object matching and grabbing capability, and determining whether the object is directly placed at a target position or not according to the matching condition of the current object, or temporarily placing the object in a buffer area; according to the method, object matching, inner ring strategy learning and implementation of an inner ring termination mechanism are carried out by utilizing a pre-trained visual basic model, the method is more robust to the influence of perception noise, and the optimality of a task level is considered through reinforcement learning closed-loop optimization, so that a better task success rate and efficiency are achieved.

A kind of high reliability anti-reverse connection circuit suitable for redundant dual controller, increase high side MOS gate auxiliary circuit and sensor low side MOS control circuit.When power is connected, auxiliary circuit does not work;When power is reversed, triode T1, T2 is turned on, and then the gate voltage of MOS Q1, Q2 is pulled to the GND of reverse connection, so that the gate-source voltage of Q1, Q2 is lower than threshold value, to ensure the reliable and stable turn-off of MOS on power bus, at the same time, the gate voltage of MOS Q3, Q4 is also pulled to the GND of reverse connection, so that the gate-source voltage of Q3, Q4 is lower than threshold value, to ensure the reliable and stable turn-off of MOS on sensor ground wire.Diode D1 and D2 realize that this part of circuit does not conduct when power is connected, and only works when power is reversed, resistance RH and RL provide base bias voltage of triode and play the role of current limiting.This application realizes the high reliability anti-reverse protection of redundant dual controller system and external common ground sensor through the distributed configuration of high and low side MOS on the input end of automobile main load power supply and the backflow end of sensor ground wire and the design of targeted auxiliary circuit, and the anti-reverse circuit design scheme suitable for redundant dual controller can effectively realize the high reliability anti-reverse protection.

The embodiment of the present specification provides a curved surface reconstruction angle selection method, system and device. The system can determine the included angle between the distance vector between the sampling points on the center line of the reconstruction object and the candidate interest vector representing the candidate observation angle, and then determine whether the candidate observation angle corresponding to the candidate interest vector is selected according to the included angle. In this way, the candidate observation angle leading to the self-crossing phenomenon can be automatically identified, and the self-crossing part in the curved surface reconstruction image can be avoided to mislead the doctor.

The application discloses a terminal crimping die tracking information management method for automobile wire harness, relates to the technical field of information processing, and comprises the following steps: acquiring crimping records generated in a continuous operation state and corresponding time position marks, performing trajectory sketching on sequence changes of the crimping records in a writing process, performing expansion processing on the crimping records according to a time advancing relationship, and forming basic record data. The application realizes identification and positioning of repeated writing and covering behavior by sketching sequence changes of crimping record writing and extracting periodic characteristics, restores a real advancing relationship of records, guarantees accuracy of a corresponding relationship of data, controls subsequent writing rhythm at the same time, makes record distribution tend to be balanced, suppresses a covering trend, maintains continuity of a record trajectory, and improves data stability and quality traceability reliability.

The present invention discloses an intelligent multi-scale registration method for cross-modal remote sensing images for complex ground scenes, and belongs to the field of aerospace remote sensing image processing and computer vision technology. On the one hand, the present invention improves the robustness of the similarity metric for cross-modal remote sensing image representation by aligning modal alignable structures and filtering structures that are difficult to align between modalities; on the other hand, by adopting a coarse-to-fine multi-level matching strategy, it effectively avoids searching for each pixel in the area, improves the search efficiency, and also avoids the missed matching phenomenon caused by the method of too strong a priori region. Through the above two aspects, the present invention achieves efficient acquisition of high-precision pixel-level matching between cross-modal remote sensing images.

The utility model provides an IGBT driving circuit based on a push-pull structure, and relates to the technical field of power electronic device driving, and the IGBT driving circuit comprises an optocoupler isolation module which is used for receiving a control signal and providing electrical isolation output; the push-pull driving module comprises an NPN type triode Q1 and a PNP type triode Q2 which are complementary and symmetrical; the push-pull driving module is configured as follows: when the optocoupler is conducted, Q1 is saturated and conducted, Q2 is cut off, and a driving current charges an IGBT grid electrode through R2; when the optocoupler is turned off, the Q2 is saturated and turned on, the Q1 is turned off, and the driving current flows through the R3 to discharge the grid electrode of the IGBT. Bidirectional driving current is provided through a push-pull structure, the charge charging and discharging speed of the IGBT grid electrode is increased, the switching loss is reduced, bidirectional driving current is provided through a switch-on resistor R2 and a switch-off resistor R3, the resistance value of the switch-on resistor R2 is selected to be small, the purpose of reducing switch-on loss Eon is achieved, the resistance value of the switch-off resistor R3 is selected to be large, and the purpose of reducing switch-off Vce is achieved.

The invention relates to the technical field of online sales service, and discloses an online sales service method based on clothing pattern design, which comprises the following steps: a customer inputs body data; obtaining a target garment purchased by the customer and the type thereof; analyzing all historical evaluation records of the customer, and judging whether a good evaluation record the same as the target garment type is included or not; if so, calculating an average value of various data in the favorable comment record according to a first strategy, and recommending the most suitable size; if not, judging whether the customer opens the historical browsing record or not. If not, screening out a target population similar to the customer body data, analyzing a good reputation record of the target population, and recommending a size; and if open, selecting a target size according to a data change trend by analyzing a historical browsing record of the customer. According to the method, accurate size recommendation is realized by comprehensively analyzing customer body data, historical evaluation records and browsing behaviors.

The invention discloses a driving circuit for a double-path output LED vehicle lamp. On and off of the two LED lamp sets are controlled by adopting an input MOSFET full-bridge rectifier circuit, a diode signal rectifier bridge and an output circuit of the LED lamp sets, the MOSFET full-bridge rectifier circuit integrates an existing MOSFET non-polarity reverse connection circuit, the circuit is simplified, meanwhile, an anti-interference first-stage triode control circuit is added, the applicability is wider, and use is more convenient; in addition, non-polarity can be achieved through the input bridge rectifier module, a user does not need to worry that a lamp is not turned on or automobile fuses are burnt out due to reverse insertion, and an original automobile is not damaged. The driving circuit is suitable for driving the double-path output LED car lamp, the LED car lamp is prevented from being damaged after input end signals are reversely connected, and meanwhile two paths of LED car lamp selection output are provided.

This invention relates to the field of natural language processing, specifically to a method and system for evaluating and selecting training value of corpus for incremental pre-training. The method involves: acquiring the forward inference results of a base model on a general corpus; segmenting the acquired candidate incremental corpus into text segments; performing forward inference on each text segment using the base model to obtain a first predicted probability distribution; performing forward inference on the lightly augmented text segments using the base model to obtain at least one second predicted probability distribution; calculating four evaluation signals for each text segment based on the first and second predicted probability distributions; weighting and fusing the four evaluation signals to obtain a comprehensive training value score for each text segment; and selecting the target corpus from the candidate incremental corpus based on the comprehensive training value scores of all text segments. This invention solves the problem of a single dimension in corpus value evaluation, effectively controls noisy samples, and significantly improves the resource utilization efficiency of incremental pre-training.

The application discloses a kind of anti-degradation multi-sensor fusion SLAM method and system, belong to robot synchronous positioning and mapping technical field.The method includes obtaining original data from camera, laser radar, inertial measurement unit and global navigation satellite system, based on VIG tight coupling optimization unit, original data is preprocessed and initial estimate in front end, obtain global pose, based on LIG tight coupling optimization unit, motion distortion correction and feature extraction processing are carried out by original data and global pose, based on two-stage factor graph optimizer, laser feature and GNSS observation information are fused in two stages, global optimization is executed, and high-precision, globally consistent pose and map are output.The application overcomes the limitation of single sensor, suppresses cumulative error, realizes high-precision, high robustness and global consistency positioning and mapping in large-scale complex scene outdoors.