108results about How to "Avoid noise disturbance" patented technology

The invention relates to a non-signaling test-based frequency offset calibration method, system and device and a medium, and the method comprises the steps: carrying out the parallel non-signaling frequency offset test of a plurality of target chips, collecting the output waveforms of the target chips, and determining the frequency offset values of the target chips according to the output waveforms; performing deviation analysis on the frequency deviation value, and calculating a corresponding frequency deviation calibration value based on the frequency deviation value and a preset frequency calibration coefficient under the condition that the frequency deviation value reaches a preset frequency limit range; writing the frequency offset calibration value into a corresponding target chip, and performing parallel frequency offset calibration; and carrying out calibration degree verification on the target chip after frequency offset calibration so as to confirm whether the frequency offset value of the target chip accords with a preset standard frequency. The method and the device have the effect of improving the frequency offset calibration efficiency of the chip.

The invention provides a ship power failure early warning and diagnosis system and a method thereof, and relates to the technical field of power measurement, a power belief network of a target ship power grid is constructed according to a topological connection relationship among different cable sections in a target ship, and association confidence among nodes in the power belief network is determined; performing measurement deviation estimation on the first power data of different cable sections, when an estimation result is judged to be abnormal, correcting the first power data through a power belief network to obtain power correction data, and determining a transaction coefficient of the target ship power grid according to the association confidence and the power correction data; and when the transaction coefficient is greater than a preset transaction threshold, predicting a fault state of each cable section according to the second power data and the power correction data to obtain a fault response value of the cable section, and performing partition diagnosis on the target ship power based on each fault response value. Therefore, the power fault diagnosis precision is improved.

The invention discloses a pneumovirus gene data multi-view clustering integration method and device and electronic equipment. The method comprises the following steps: constructing a width self-encoder with a width learning network and a global subspace structure; training a width auto-encoder, determining model parameters of the width auto-encoder, inputting multi-view pneumonia gene data into the trained width auto-encoder, and outputting a compressed feature subspace; clustering the compressed feature subspaces to obtain a plurality of basic partitions as integration members of an integration pool; the basic partition is converted into a fuzzy partition matrix, an undirected bipartite graph is constructed based on the integration pool, and connection weights between nodes of the undirected bipartite graph correspond to values in the fuzzy partition matrix; and cutting the undirected bipartite graph into several disjoint sub-graphs, and obtaining a clustering integration result of multi-view pneumonia gene data according to the sample nodes of each sub-graph. According to the method, parameter quantity and computing resources are remarkably reduced, and multi-view clustering analysis of pneumonia gene data is efficiently completed.

The invention provides a non-invasive multi-frequency dessert displacement monitoring method, which comprises the following steps of: setting an excitation frequency set and a frequency switching interval through an upper display control machine, and controlling a lower embedded system to switch different frequencies in real time according to the set interval, thereby realizing real-time switching of multiple frequency points; the lower embedded system collects electrocardiosignals and cardiac impedance signals of a monitored object and transmits the electrocardiosignals and the cardiac impedance signals to the upper display and control machine through wired or wireless WIFI / Bluetooth. And the upper display control machine performs corresponding filtering and wavelet transform analysis on the cardiac impedance signal and the electrocardiosignal, extracts important feature points of the signals and calculates a cardiac displacement index. The lower embedded system is controlled by the upper display and control machine to monitor multi-frequency excitation of a human body, interference of different-frequency noise on measured electrocardio and cardiac impedance signals in different environments is avoided, and non-intrusive continuous cardiovascular monitoring can be realized. And more accurate hemodynamic parameters are provided for clinical diagnosis and treatment.

The invention belongs to the technical field of network security, and discloses an enterprise network security self-checking method and system, and the method comprises the steps: collecting network flow data, equipment parameter data, user behavior data and security situation awareness data; preprocessing the collected network traffic data, equipment parameter data and user behavior data to obtain a network traffic feature data set, an equipment feature data set, a user behavior feature data set and a security situation awareness feature data set; fusing the network traffic feature data set, the equipment feature data set, the user behavior feature data set and the security situation awareness feature data set to obtain a comprehensive feature data set; training according to the comprehensive feature data set to obtain a network security prediction model, and predicting through the network security prediction model to obtain an enterprise network security state; the possibility of occurrence of network security events is reduced, response delay caused by human intervention is reduced through an automatic process, and the detection efficiency is greatly improved.

The invention discloses an operation and maintenance knowledge graph construction method, a track operation and maintenance method and a related device, parameter initialization of a BiRTE model is adopted when a PRGC model is trained, so that the PRGC model has the stability of the BiRTE model in relation extraction, the accuracy of operation and maintenance knowledge triple extraction is ensured, and the operation and maintenance knowledge graph construction method is suitable for the track operation and maintenance. A multi-task learning model is adopted to filter noisy operation and maintenance knowledge triads, noise interference existing in the operation and maintenance knowledge triads is solved, and therefore it is guaranteed that the accuracy and robustness of the constructed operation and maintenance knowledge graph are good.

The invention discloses a liver cancer post-transplantation muscle loss prediction method and device, and belongs to the technical field of biomedicine, and the method comprises the steps: constructing an unbalanced random forest model comprising a plurality of decision trees, independently predicting through the plurality of decision trees to obtain a plurality of high-risk or low-risk dichotomy prediction results, taking the proportion of the decision tree with the prediction result being high risk in all decision trees as the prediction probability of muscle loss after liver cancer liver transplantation finally output by the model; performing risk classification according to the prediction probability, and judging the postoperative muscle loss level of the patient as high risk, medium risk or low risk; constructing a decision tree model to layer prognosis of patients, and dividing the patients into a high-risk group, a medium-risk group and a low-risk group; and carrying out survival rate prediction in combination with liver transplantation standards and prognosis layering results of patients. According to the method, the muscle loss level after the liver transplantation can be accurately predicted, and a clinician can make a personalized treatment scheme and perform scientific management after the transplantation.

The embodiment of the invention provides an intestinal polyp segmentation method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring an intestinal polyp image; preprocessing the intestinal polyp image to obtain a standardized intestinal polyp image; and feature extraction is carried out on the standardized intestinal polyp image, multiple image features of different scales are obtained and input to the intestinal polyp segmentation model, an intestinal polyp segmentation result is obtained, and the wavelet-space high-order feature aggregation network comprises a boundary learning module, a multi-scale feature learning module and a multi-scale feature aggregation module. The boundary learning module is used for learning boundary features of intestinal polyps, the multi-scale feature learning module is used for learning image features of multiple different scales to obtain supplementary features, and the multi-scale feature aggregation module is used for carrying out feature fusion on the image features of multiple different scales, the boundary features and the supplementary features to output a segmentation image. Based on this, the embodiment of the invention can realize more accurate and efficient intestinal polyp segmentation.

The invention discloses a neighbor entity-based medical knowledge graph automatic alignment method. The method comprises the following steps: 1) acquiring a source medical knowledge graph and a target medical knowledge graph; 2) obtaining a basic vector representation of the medical entity by using the pre-training model; 3) constructing a neighbor relation mask matrix; 4) constructing a neighbor entity interaction matrix; 5) obtaining the comprehensive similarity between the entities and finely adjusting the model by using a loss function; and 6) checking the alignment degree between the medical knowledge maps by using the entity alignment evaluation indexes. According to the medical knowledge graph automatic alignment method based on the neighbor entities, the model considers the influence of the similarity between the entities on entity alignment between two medical knowledge graphs with heterogeneity, and also considers the influence of the similarity between the neighbor entities of the entities on the entity alignment effect; and the generalization ability of the model is enhanced, so that contributions are made in the fields of knowledge graph fusion, data governance, text understanding, question and answer systems and the like.

The invention relates to the field of material sampling and testing, in particular to a drilling and coring device for bridge pile foundation detection and a control method of the drilling and coring device. An infrared distance meter is mounted on a vertical rod of a sliding block on an annular guide rail on the base; the sampling barrel penetrates through the sleeve, the bottom end of the sleeve is connected with the base, a hydraulic cylinder of the hydraulic equipment is hinged to the base, a blocking piece hinged to the top end of a piston rod of the hydraulic equipment abuts against the top end of the sleeve, an annular groove is formed in the outer wall of the top end of the sleeve, and a pressure sensor is installed between the blocking piece and the annular groove. A processor is mounted in a box body on the base; in the drilling and coring process, the sliding block moves along the annular guide rail, the processor collects data output by the infrared distance meter and the pressure sensor and obtains the deviation direction and the deviation angle of the sampling barrel, and the pressure of the hydraulic equipment is adjusted according to the deviation direction and the deviation angle. The problem that the sampling barrel cannot automatically rectify deviation in real time when the solid material to be analyzed is sampled, drilled and cut is solved.

The invention relates to a method and system for detecting abnormal deflection of a tool. The method comprises the steps that first vibration signal data generated in the idling process of a tool spindle are collected through a vibration sensor, the first vibration signal data are processed, and training data are obtained; based on the training data, training a class of SVM models to obtain a trained classification model; second vibration signal data of the tool spindle in the real-time operation state are collected through a vibration sensor, the second vibration signal data are processed, and the processed second vibration signal data are obtained; on the basis of a classification model, predicting the processed second vibration signal data to obtain a classification result, and on the basis of the classification result, obtaining a proportion of a positive sample to a negative sample; calculating the deviation degree of the proportion of the positive sample and the negative sample in the classification result; the deviation degree is compared with a preset threshold value, and when the deviation degree is larger than the threshold value, the NC system sends out an alarm signal. And the abnormal deflection detection efficiency and the response speed of the cutter are improved.

The invention discloses a fish school identification and classification method in a complex environment based on a particle swarm and improved YOLOv6. The method comprises the following steps: firstly, carrying out image enhancement processing on an acquired fuzzy light image by applying a multi-scale Retinex algorithm on an image acquisition level; secondly, an improved BC-YOLOv6 algorithm is applied to the image recognition level, Backbone of YOLOv6 is combined with BiFormer of a Transform layer of a bidirectional structure, a coordinate attention mechanism module is introduced into the architecture of the BC-YOLOv6 algorithm, the recognition rate of small targets is increased, non-maximum suppression is improved, and correlation functions are designed; and finally, defining a feature matrix and improving a particle swarm algorithm on an image classification level, and obtaining the selected feature matrix to realize image classification. Simulation results show that manpower and computing power cost can be effectively reduced, enhancement processing on the images is remarkably improved, and the monitoring rate, the accuracy, the feature expression ability and the image classification performance of underwater fish school image recognition and classification under the shielding condition are further improved.

The invention belongs to the technical field of weld joint detection, and particularly relates to a weld joint defect detection method and device. The method comprises the following steps: S1, generating a magnetic field at a weld joint to be detected by using an excitation probe; s2, original magnetic field signals of the magnetic field in the X axial direction, the Y axial direction and the Z axial direction are collected; s3, processing the original magnetic field signal in each axial direction by using wavelet transform to obtain characteristic magnetic field signals Bx, By and Bz containing defect characteristics in three axial directions; s4, respectively calculating amplitude difference values between wave crests and wave troughs in the oscillogram of the characteristic magnetic field signals Bx, By and Bz; and if at least one of the amplitude difference values in the three oscillograms is greater than a preset threshold value, determining that the weld joint has defects. Through the excitation probe and a magnetic field signal processing method matched with the excitation probe, noise interference can be removed to accurately detect defects and estimate reference sizes of the defects, so that the purpose of improving the detection accuracy is achieved.

The invention relates to the technical field of particle detection, in particular to a SiPM test and calibration method and system for particle detection application, and the method comprises the following steps: extracting a slope quadratic difference set to determine a current change inflection point, calculating a local slope change proportion set to obtain a breakdown breakover voltage point, analyzing the output pulse triggering probability, and calculating the current change inflection point; and in combination with a photon number extraction probability growth interval, constructing a refined pulse extraction detection efficiency reference ratio, screening high-density trigger data, and calculating a spatial deviation extreme value, so that the system bias compensation voltage is adjusted. According to the method, the breakdown breakover voltage point is positioned, so that measurement interference caused by basic background noise is avoided, the attenuation number is counted to establish a trigger probability sequence, a preset maximum likelihood estimation model is combined to extract a detection efficiency reference ratio, a channel deviation extreme value is calculated to directly correct parameter fluctuation, and array global dynamic balance is established. And the long-term stability of the whole detection working state is maintained.

The invention discloses a code abstract generation method and system based on hierarchical context awareness, and belongs to the field of natural language processing and software engineering. The method comprises the following steps: generating hierarchical annotations for a code warehouse based on a function call graph, firstly extracting a call relationship in the graph and eliminating a ring structure, and then classifying leaf nodes and generating annotations; a topological sequence from bottom to top is adopted in the core step, the importance score of a child node is calculated for each father node, child node annotations, class function descriptions and code bodies of the child nodes are fused, and a large language model is input to generate annotations till the whole graph is covered. The method is suitable for an automatic document generation scene of a complex software project, and through a technical path combining function call graph analysis, child node importance measurement and class function description, the accuracy and semantic consistency of code annotation are improved; the method can be widely applied to the fields of software maintenance, code understanding, intelligent development tools, automatic document generation of enterprise-level code warehouses and the like.

The application discloses a colorectal cancer metastasis evaluation method based on NF-kappa B pathway activity, relates to the technical field of medical treatment, and comprises the following steps: obtaining a colorectal cancer tissue sample and a paired peripheral blood sample, separating exosomes, and dividing the tissue sample into a spatial resolution sub-sample and a function verification sub-sample; performing multiple immunofluorescence imaging on the spatial resolution sub-sample, identifying a phosphorylated p65 positive cancer cell group and a phosphorylated STAT3 positive immune cell group, and extracting cell phenotype data; calculating the spatial interaction weight value between high-activity cells based on the cell phenotype data to obtain a tumor microenvironment paracrine activation index; detecting the abundance of an upstream phosphorylated protein of the NF-kappa B pathway and the expression amount of a downstream metastasis effect gene in the function verification sub-sample, and obtaining a pathway function coupling score in combination with the matching degree of the two; and determining the content of a phosphorylated IKK protein and the content of a specific metastasis-promoting microRNA in the exosomes, and obtaining an exosome double-lock activity score by using the product relationship between the two.

The invention belongs to the field of image processing, and discloses an infrared and visible light image fusion method based on cross-domain feature fusion and GAN, and the method comprises the steps: effectively extracting the feature information of a shallow layer and a deep layer in a source image through designing a shallow layer-deep layer feature extraction module, and reducing model network parameters; by constructing a cross-domain feature fusion module, cross-modal complementary information is mined and integrated, information loss and noise interference during image fusion are reduced, and the representation intensity of a fused image is ensured; by designing a main and auxiliary feature loss function, structure information, texture information, gradient information and secondary information in a source image are fully extracted, and finally a model is trained to obtain a detection result and perform evaluation. According to the method, while the contrast and texture details of the source image are fully reserved, hidden auxiliary information in the source image can be deeply mined and effectively utilized, and the overall quality of the fused image is remarkably improved.

The embodiment of the invention provides a vehicle-mounted engine power generation control method and device, electronic equipment and a storage medium, and the method comprises the steps: judging whether the residual electric quantity supports the pure electric driving of a vehicle to a destination or not according to the current travel information, and if not, generating a driving signal; acquiring a driving speed and a running state of a vehicle-mounted player in response to the driving signal; controlling a vehicle-mounted engine to start based on the residual electric quantity, the running speed and the running state of the vehicle-mounted player; and adjusting the operation power of the vehicle-mounted engine according to the volume intensity corresponding to the operation state of the vehicle-mounted player, so that the volume of the vehicle-mounted player covers the operation noise of the vehicle-mounted engine. According to the technical scheme, on the premise that the vehicle-mounted engine provides reasonable electric energy for the electric vehicle, interference of noise generated by the vehicle-mounted engine to a driver and passengers is reduced, and the driving experience feeling of the driver and the passengers is improved.

The application discloses a kind of pathological image classification methods based on dyeing perception intelligent feature modeling, belong to pathological image classification field, this method includes collecting the multi-staining full section image group data of different objects;Multi-staining full section image group data includes H&E dyeing image and several immunohistochemical dyeing images;Each multi-staining full section image group data is respectively preprocessed, obtains the several equal-size image blocks of each object;Based on the image block of each object, graph is constructed, and the multi-modal heterogeneous pathology graph of each object is obtained;With the multi-modal heterogeneous pathology graph of each object as training data, train pathological image classification model;The pathological image classification model of training completion is used to classify the pathological image to be measured.The application solves the problems of the existing method in processing digital pathology graph data, such as node category characteristics, edge information utilization, cross-staining interaction and non-perfect data processing.

The invention belongs to the technical field of accelerometer second-order system identification, and particularly discloses an accelerometer system online parameter identification method based on a digital phase-locked loop, which comprises the following steps: in a first displacement signal detection stage, applying a first excitation signal to a mass block in an accelerometer through electromagnetic excitation, and detecting a first displacement signal of the mass block, the phase difference between the first displacement signal and the first excitation signal is determined as a reference phase difference, and the frequency of the excitation signal in the first displacement signal detection stage is fixed in a preset frequency range; in the second displacement signal detection stage, a second displacement signal of the mass block is detected, the target phase difference is tracked through the digital phase-locked loop, the difference between the target phase difference and the reference phase difference is # imgabs0 #, and the frequency of the excitation signal in the second displacement signal detection stage is larger than the preset frequency upper limit; and after the phase-locked loop is stable, determining the eigenfrequency and the quality factor of the accelerometer. According to the invention, rapid online parameter identification of the accelerometer system is realized.

The application discloses a Mark point coordinate detection method based on an image sequence and a related device, and is used for improving display screen Mark point positioning precision. Image acquisition of N frames in succession is performed on a to-be-positioned object provided with a Mark point; Mark point coordinates of a photographed image sequence are extracted to generate a candidate coordinate point set; according to each candidate coordinate point, a gray feature of a photographed image where the candidate coordinate point is located is extracted to generate a gray feature set; a position deviation amount of each candidate coordinate point is calculated according to the candidate coordinate point set to generate a position deviation amount set; a gray consistency parameter of each candidate coordinate point is calculated according to the gray feature set; a multi-dimension confidence index of each candidate coordinate point in the candidate coordinate point set is calculated according to the position deviation amount set and the gray consistency parameter set; the candidate coordinate point set is screened according to the multi-dimension confidence index set, and a target coordinate of the Mark point is generated by using a screened candidate coordinate point subset.

The present invention relates to the field of node data processing, and in particular to a distributed data center node optimization method and system, the method comprising: collecting data of each node to obtain a data sequence of each node; calculating the influence of each node; obtaining a comprehensive stability coefficient of each node according to the maximum value, minimum value and standard deviation of the data sequence of each node; obtaining each IMF component; calculating the data fluctuation factor and local stability factor of each IMF component sub-segment; obtaining the decomposition variation factor of each IMF component according to the relationship between each IMF component and the corresponding node data sequence; obtaining the local stability coefficient of the node according to the local stability factor and decomposition variation factor of each IMF component; and then obtaining the stability index of each node; obtaining the importance of each node in combination with the influence and stability index of each node, selecting a data center node according to the importance, and completing the optimization processing of the data center node. Thus, efficient selection of data center nodes is achieved.

The invention discloses a high-capacity circuit breaker heat dissipation structure based on a micro-nano structure heat pipe, and belongs to the technical field of circuit breaker heat dissipation. Comprising a vacuum arc extinguish chamber, an arc extinguish chamber heat dissipation device, a micro-nano structure heat pipe and a heat pipe heat dissipation device. The arc extinguish chamber heat dissipation devices are arranged on the two sides of the vacuum arc extinguish chamber, the micro-nano structure heat pipe is embedded in a static conducting rod of the vacuum arc extinguish chamber, and the heat pipe heat dissipation fins are arranged on the heat pipe. Through cooperation of the arc extinguish chamber heat dissipation device and the micro-nano structure heat pipe, radial conduction and axial rapid conduction of heat of the vacuum arc extinguish chamber are realized, and temperature rise of the high-capacity circuit breaker is remarkably reduced.

The application provides a Chinese-oriented named entity recognition method and system, and belongs to the technical field of natural language processing; comprising: an encoding stage step, context encoding of input Chinese text; a feature extraction stage, a double-branch structure of global semantic branch and local boundary branch in parallel is adopted, wherein the local boundary branch carries out boundary feature enhancement through multi-granularity convolution combined with a learnable Gaussian Laplacian operator; a feature fusion and prediction stage step, global and local features are fused through a gating mechanism, and multi-label classification prediction is carried out based on a preset six-tuple grid label system; a decoding stage step, named entities are obtained through bidirectional search strategy decoding according to predicted relationship probability; through boundary perception feature extraction and enhanced label system, the recognition accuracy and recall rate of nested entities and discontinuous entities in Chinese text are significantly improved.

The invention discloses a rotary water-leakage-proof aluminum alloy door and window, and relates to the technical field of aluminum alloy doors and windows, in particular to a rotary water-leakage-proof aluminum alloy door and window which comprises a wall body, a servo motor is fixedly installed on the right side of the front face of the wall body, and a first driving wheel is fixedly installed on an output shaft of the servo motor. According to the rotary water-leakage-proof aluminum alloy door and window, a first driving wheel can be driven to rotate by starting a servo motor, a first crawler belt can be driven through rotation of the first driving wheel, so that a first driven wheel rotates, then a rotating shaft can be driven, and a second driving wheel rotates; a second driving wheel rotates to drive a second caterpillar band so that a third driving wheel can rotate, the second caterpillar band can drive a first driving block and a supporting keel at the same time, the sunshade curtain can be rapidly opened, closed and folded, and meanwhile the noise trouble caused by a traditional sunshade curtain driving chain is avoided.

The application discloses a kind of measuring circuit and test equipment, measuring circuit is connected with the device under test, for performing test to electrical signal from device under test, measuring circuit includes source measurement module and preamplifier, preamplifier includes operational amplifier, first conversion module and second conversion module;Source measurement module and the noninverting input of operational amplifier are connected, the noninverting input and the output of operational amplifier are connected with first conversion module, the output of operational amplifier is connected with device under test by second conversion module, the inverting input of operational amplifier is connected with device under test, source measurement module outputs first main signal, obtains second main signal by transmission and is transmitted to device under test;And source measurement module samples and measures current signal in first main signal according to current signal value and the current gain of preamplifier, obtains the actual current value flowing through device under test.

The invention relates to the technical field of sensors, in particular to a multi-dimensional force sensor which comprises a shell, an elastic body, a supporting structure and a plurality of dampers. The elastic body is arranged in the shell and comprises a load input end for receiving an external load and a mounting part connected with the shell, and the elastic body is used for generating deformation in response to the external load; the supporting structure is arranged on the periphery of the elastic body in a sleeving mode and located in the shell, and the supporting structure is connected with the load input end; the multiple dampers are arranged between the supporting structure and the shell; wherein the damper is configured to dissipate the vibration energy of the elastic body through a viscous dissipation effect or flexible deformation. According to the multi-dimensional force sensor provided by the invention, the dynamic performance can be directly and effectively improved from a mechanical structure, the self vibration is inhibited, and the response time is shortened.

The application provides a kind of tool notch detection method, device, equipment and readable storage medium, applied to defect detection technical field.The edge of rotating tool is detected in real time by using photoelectric sensor, wherein the edge of tool at least shields part of detection light path of photoelectric sensor, and tool detection signal output by photoelectric sensor is acquired in real time.The light flux of detection light path not being shielded by tool is positively correlated with tool detection signal output by photoelectric sensor, and the time corresponding to one rotation of tool is used as a period, the maximum extreme point and minimum extreme point in tool detection signal waveform are found in each period, and finally whether tool exists notch is determined according to the slope of line between maximum extreme point and minimum extreme point.It can be seen that the tool in rotating working state can be detected in real time by photoelectric sensor, whether tool exists notch can be quickly judged, and the accuracy of tool damage notch detection result is improved.

This application discloses a method, apparatus, and equipment for adjusting navigation tasks based on human intelligence function allocation, belonging to the field of ship control technology. The method includes: performing cross-band dynamic functional connectivity analysis on physiological signals to extract neural state trajectories; performing task semantic parsing on navigation operation logs to extract behavioral log streams corresponding to the ship's sub-task types; constructing positive and negative sample pairs based on the synchronization state of the neural state trajectories and behavioral log streams; mapping the neural state trajectories, behavioral log streams, and environmental pressure vectors to a maritime context-aware latent space with the optimization objectives of maximizing the mutual information of positive sample pairs, minimizing the distance of negative sample pairs, and maximizing the correlation between negative sample pairs and the environmental pressure vector, thereby obtaining a comprehensive contextualized common representation of the crew; and calculating metacognitive alignment loss based on the comprehensive contextualized common representation to adjust the crew's navigation tasks. This method can improve the contextual sensitivity of task adjustment.

The present application provides an asphalt pavement apparent disease evolution deduction method, system, device and medium, which belongs to the field of road engineering, transportation infrastructure operation and maintenance and intelligent detection technology, and comprises the following steps: constructing a unified space-time reference system, aligning and normalizing multi-source data in space-time; forming section-level disease quantification characteristics based on the recognition results of the inspection images and the structural indexes; event coding the maintenance measures of the road surface and pre-processing the section-level disease quantification characteristics corresponding to the time of the maintenance measures; constructing the correlation, time lag effect and spatial consistency between the disease types, disease characteristics and structural indexes to generate the collaborative evolution characteristics of the coupling relationship between diseases; constructing a deep time series prediction model to obtain the disease development trend at a specified time scale in the future, and forming the section-level disease evolution prediction result. The present application can depict the disease development law at different road sections and different time scales, and predict the future disease expansion.