760results about How to "Increased complexity" patented technology

The invention discloses an intelligent customer service method for adaptively adjusting an interaction strategy by using reinforcement learning, and the method comprises the following steps: S1, obtaining a natural language text inputted by a user, and generating a multi-dimensional state vector; s2, grouping the users by using a real-time dynamic clustering algorithm, and updating a clustering result; s3, generating an action space in a specific state, and enriching the action space in combination with a rule-based strategy and a learning-based strategy; s4, aggregating neighbor node features through a graph convolutional network, and extracting local features; s5, weighting adjacent node features by using a graph attention network, and dynamically allocating attention weights; s6, carrying out interaction strategy optimization by adopting a spiral ladder type deep exploration network, and optimizing action selection; and S7, combining online learning with an experience playback mechanism, continuously updating an interaction strategy, and making a quick response to dynamically changing user demands. According to the method, reinforcement learning, the graph neural network and the sentiment analysis technology are combined, and intelligent customer service interaction strategy self-adaptive adjustment and optimization are achieved.

The invention discloses a text-driven automatic image generation method and system, and belongs to the technical field of image generation, and the generation method comprises the following specific steps: I, receiving and analyzing text data input by a user, carrying out explicit modeling on entities and relationships in a text, and extracting implicit languages in the text data; according to the method, the multi-modal alignment of the text and the visual features is improved, the visual deviation problem caused by semantic information errors is avoided, the detail richness and semantic information expansion capability of the generated image are improved, the global layout reasonability of the generated image is improved, and the individuation and stylization capabilities of the generated image are enhanced; according to the method, a rigid effect caused by a fixed template is avoided, the performance of the generated image on scene complexity and detail richness is improved, the accuracy of relation modeling between objects in a complex scene is remarkably improved, the intelligent level of image generation is improved, and the generation process is more adaptive.

The invention discloses a method and system for improving satellite remote sensing inversion precision by fusing hyperspectral proximity, and belongs to the technical field of water environment and water ecology monitoring. Acquiring a water body hyperspectral near-sensing reflectivity spectral data set, a satellite remote sensing surface reflectance data set and a water environment actual measurement data set, and performing preprocessing; a synchronous data set containing # imgabs0 # and water quality and a synchronous data set containing # imgabs1 # and water quality are constructed; based on the synchronous data of the # imgabs2 # and the water quality, synchronous data of the # imgabs3 # and the water quality are generated, and synchronous data of the # imgabs4 # and the # imgabs5 # are obtained; the # imgabs6 # is normalized to obtain # imgabs7 #, and the quantitative conversion relation between the # imgabs7 # and the # imgabs8 # is constructed band by band to eliminate errors; generating a fused satellite-ground data set, and constructing a training water quality parameter inversion model; and updating corresponding methods and data sets in real time by repeating specified steps, and optimizing the model. According to the method, the inversion result is displayed and downloaded according to the query instruction, the satellite spectrum precision and the satellite-ground synchronous data set scale are improved, and the method is used for high-precision water environment monitoring.

The invention relates to a method for evaluating the situation awareness ability of a pilot based on a Ti-SA model. The Ti-SA model comprises an eye movement trace multi-dimensional attention enhancement module, an AOI multi-scale comprehensive feature module and a fusion module. The evaluation method comprises the steps of collecting eye movement trace time sequence data and flight training videos of a pilot at continuous moments in real time, and obtaining corresponding AOI fixation time sequence data based on the eye movement trace time sequence data, the flight training videos and a preset AOI subdivision area; preprocessing the eye movement trace time sequence data and the AOI fixation time sequence data; inputting the preprocessed eye movement trace time sequence data into an eye movement trace multi-dimensional attention enhancement module to obtain a category probability value corresponding to the eye movement trace time sequence data; inputting the preprocessed AOI fixation time sequence data into an AOI multi-scale comprehensive feature module to obtain a category probability value corresponding to an AOI fusion feature; and inputting the two category probability values into a fusion module to obtain a predicted situation awareness level.

The invention relates to the technical field of hydraulic engineering, geotechnical mechanics and geological disaster prevention and control engineering, and discloses a landslide surge physical and numerical hybrid model construction method and system and application. Analyzing the landslide instability evolution motion process, the landslide warehousing first wave height and the surge propagation process of the landslide surge; performing a numerical simulation test according to a second result of the landslide surge numerical model test; adjusting the particle size and the kernel function in the landslide surge numerical simulation test, so that the first wave height, the opposite-bank arrival height and the surge pressure distribution in the landslide surge numerical simulation test and the landslide surge physical model test are consistent; and taking the adjusted landslide surge numerical model as the obtained landslide surge hybrid model. The system comprises a first result acquisition module, a second result acquisition module, a test parameter adjustment module and a hybrid model acquisition module. According to the method, different calculation working conditions can be conveniently analyzed.

The invention discloses a chronic disease medical and anti-medical fusion collaborative management and control system, belongs to the technical field of health data processing, and solves the problem that when an existing system extracts an electronic case file of a user, the health condition change of a patient cannot be reflected in real time, and consequently key feature extraction is inaccurate. The system comprises a user side, a fusion cloud platform, a medical defense judgment module and a follow-up visit customization module, and the fusion cloud platform is used for executing a risk assessment model and outputting a risk assessment result by taking a normalized data set associated with a user as input; a risk assessment model based on deep learning and an attention mechanism is constructed by fusing a cloud platform, so that the risk assessment model can identify and analyze a normalized data set collected by a user side, and the user side can collect a multi-source heterogeneous normalized data set which is good in timeliness and reflects the health condition of a patient in real time; therefore, it is ensured that the risk assessment model provides an accurate assessment path for chronic disease medical prevention and control.

The invention discloses a tunnel drainage pipeline blockage condition monitoring and processing method and system, and the method comprises the steps: collecting data, such as optical images, three-dimensional point cloud, ultrasonic reflection, gas components, infrared imaging, sound wave images, electromagnetic induction and flow velocity, through a multi-mode sensor array, carrying out the denoising and normalization preprocessing, and extracting multi-scale features; fusing and analyzing the feature data by using a self-attention mechanism and a graph neural network to generate optimized multi-modal comprehensive feature data; based on the data, a full-connection neural network discriminates the type, the number, the size, the blockage degree and the cleaning difficulty of blockages, and dynamic correction is carried out through real-time data feedback and a self-learning mechanism; and finally, cleaning scheme data is generated, and a cleaning task is executed through the mechanical arm. According to the method, the comprehensiveness of data acquisition, the depth of feature extraction, the intelligentization of feature fusion and the accuracy and the real-time performance of a discrimination model are improved, and comprehensive monitoring and efficient processing of the blockage condition of the tunnel drainage pipeline are realized.

The invention relates to an electric power virtual simulation optimization method based on a digital twin technology, and the method comprises the following steps: S1, deploying sensors at key nodes of an electric power system, and integrating an intelligent electric meter, weather forecast, social media information and market data to form a comprehensive data pool; s2, constructing a digital twinborn model of the power system; s3, optimizing the digital twin physical model based on a Bayesian optimization method and an adversarial neural network; s4, applying a bee colony algorithm to formulate a cooperation strategy, and optimizing transmission and load adjustment of power in different areas; s5, introducing an adaptive neural fuzzy inference system, and further optimizing a cooperation strategy; and S6, based on the optimized digital twinborn model and the operation strategy, performing large-scale distributed simulation by using cloud resources. According to the invention, accurate construction of the electric power simulation model and accurate implementation of electric power virtual simulation are realized.

The invention relates to the field of medical model training, and particularly discloses a medical treatment guide model training method and system, a terminal and a medium, and the method comprises the steps: carrying out the incremental pre-training of a selected large language model through priori knowledge, and obtaining a basic large language model; acquiring a public real doctor-patient dialogue data set; constructing a doctor-patient dialogue template, and generating and constructing a doctor-patient dialogue data set by using a third-party large language model based on the doctor-patient dialogue template; combining the real doctor-patient dialogue data set and the constructed doctor-patient dialogue data set into a data set, and recording the data set as an initial doctor-patient dialogue data set; performing duplicate removal preprocessing on the initial doctor-patient dialogue data set to obtain a doctor-patient dialogue training data set; and using the doctor-patient dialogue training data set to carry out supervision fine tuning training on the basic large language model to obtain a doctor-seeing and hospital-guiding model. According to the invention, the joint training data set is formed based on the public data set and the automatic construction data set, so that the accuracy of the medical treatment guide model is improved.

The invention discloses a fire-proof gas concentration detection device for a coal mine goaf, and relates to the technical field of environment detection, the fire-proof gas concentration detection device comprises a cylinder, the side wall of which is provided with an air inlet for introducing to-be-detected air; the air outlet cylinder is connected to the upper end of the cylinder body, fan blades are rotationally arranged in the air outlet cylinder, and rotation of the fan blades drives air to enter from the air inlet and flow through the whole device; the funnel-shaped dust collection cylinder is connected to the lower end of the cylinder body, and a plurality of separation blades are distributed on the inner circumference of the funnel-shaped dust collection cylinder and used for preliminarily filtering and blocking dust impurities in the air, so that the purified air enters the cylinder body to be detected; the dust collection box is connected to the bottom of the dust collection barrel; the plurality of sensor cylinders are uniformly distributed on the inner circumference of the cylinder body, a connecting seat is arranged in each sensor cylinder, and a gas semiconductor sensor is mounted on each connecting seat; the device can improve the precision of gas concentration detection, flexibly adapt to different working conditions and realize intelligent cleaning and maintenance.

The invention relates to the technical field of electric power data acquisition with an encryption function, and discloses an electric power data acquisition terminal with an encryption function and a method. By combining the environment data and the equipment state, the encryption process enhances the security and adaptability of data encryption. Environment and equipment data input ensures that encryption results are different, and potential safety hazards caused by static data sources are avoided. The data are reasonably combined through a seed generation formula, it is ensured that the seeds have enough randomness, and the brute force cracking risk is reduced. The key generation strategy generates a unique key through a seed, so that the attack resistance of the system is improved. The key matrix and random number generation strategy improves the unpredictability of the encryption algorithm. And dynamically updating the encryption matrix to prevent the encryption matrix from being cracked in long-term use. In the whole process, comprehensive protection of the data in acquisition, encryption and transmission is ensured, and the data is effectively prevented from being tampered or stolen.

The invention discloses a long time sequence prediction method based on a generative adversarial space-time attention network, and relates to the technical field of deep learning and time sequence prediction, the generative adversarial space-time attention network comprises a generative adversarial network and a space-time attention network, and the training process comprises the following steps: obtaining a time sequence data set and carrying out deletion processing; constructing a generative adversarial network, training the generative adversarial network by using the time sequence data after missing processing, the generative adversarial network being used for recovering the time sequence data after missing processing; preprocessing the recovered time series data set, and obtaining a training set and a test set; a space-time attention network is constructed, the training set and the test set are used to train the space-time attention network, and a space-time attention network model is used to predict the recovered time series data; and performing error analysis on a predicted result and an original time sequence data set. The method aims at solving the problem of missing values of the time sequence data, and meanwhile the limitation of a time sequence prediction model in the process of processing complex, high-dimensional and nonlinear time sequence data is solved.

The invention belongs to the technical field of particle motion track recognition, and particularly relates to a method for tracking a large number of moving particles based on image recognition, the moving particles suitable for the method are particles undergoing active Brownian motion, the same color mark is made on each moving particle, and image acquisition is performed through a microscope or a video recorder; identifying a pixel point cluster corresponding to the color mark on each moving particle by setting an RGB value; determining moving particles corresponding to the color marks by setting a threshold value in the pixel point cluster; taking a pixel point corresponding to the average value position of all the position data as the position of the moving particle; and recording the motion trail of the moving particles in a period of time. According to the method, the marks are introduced to the particles, the algorithm is developed to identify the marks, calculate and record the average positions of the pixel points corresponding to the marks, the particles are tracked, the colored marks can exist stably and do not interfere with one another, and it can be ensured that a large number of particles can be tracked for a long time.

The invention discloses a cultivation area prediction method based on an improved nested UNet neural network, and belongs to the field of mariculture area prediction methods. According to the method, data preprocessing is carried out on high-resolution third radar image data and high-resolution second optical image data, wave band synthesis and image fusion are carried out on the optical image data and the radar image data, and a data set is constructed. On the basis of applying a U2-Net neural network model main body structure, a model network structure of an RSU module is modified, channel attention mechanisms based on maximum pooling and average pooling are added inside and outside a residual U-shaped module structure, a U2-Net SEOUT model is constructed, and prediction of the area and distribution of a breeding area is realized. According to the method, the advantages of the optical image and the SAR image are comprehensively utilized, the network structure is modified, the number of network layers and complexity are increased, and the prediction precision of the raft cultivation area is improved.

The invention discloses a mode selection method and device for a D2D auxiliary power internet of things, and the method comprises the steps: building a graph neural network model, modeling the interference relation between D2D user pairs and cellular users in a power network and channels into a graph structure in a graph building module, coding the node features and edge features in the graph structure into feature vectors, and carrying out the coding of the feature vectors, in a graph convolution module, a graph convolution neural network is utilized to perform feature extraction on the feature vectors to obtain node hidden states including D2D user pair nodes and cellular user nodes, and in a classification decision module, the node hidden states are input into a classifier to calculate channel selection variables so as to determine a selected channel mode; and training the graph neural network model and then deploying and applying. According to the method, the pattern selection of D2D communication is optimized and the system throughput is maximized by using the graph neural network, so that the calculation complexity is further reduced, the calculation speed is improved, and the overall performance of the communication system can be greatly improved.

The invention discloses a deep learning-driven intelligent analysis method for automobile supply chain bill data, and relates to the technical field of data management.The deep learning-driven intelligent analysis method comprises the steps that an automobile supply chain bill is scanned to obtain scanned images of the automobile supply chain bill, and then the multiple scanned images obtained through scanning are preprocessed periodically to obtain a ciphertext conversion sequence; converting the ciphertext conversion sequence, and determining an array ciphertext interval of each pre-processing period according to the number of pre-processing times and the number characteristic of the segmentation array of the currently pre-processed scanning image in the process of pre-processing the scanning image, according to an array ciphertext interval, 16 kinds of originally existing four-bit binary numbers are compressed into 8 kinds, and ciphertext characters are used for increasing fault tolerance of a ciphertext conversion sequence after forward transmission, so that the ciphertext conversion sequence obtained through final conversion can have 16 kinds of four-bit binary numbers on the whole or in groups, complexity and difficulty are increased for three-party cracking, and the conversion efficiency of the ciphertext conversion sequence is improved. And the security of the scanned image in transmission is further ensured.

The invention relates to the technical field of computer vision and deep learning, and relates to a deep forgery detection method based on mask supervision. According to the technical scheme, an image is input to a preprocessing module for preprocessing; the processed image is respectively input into a mask supervision module for extracting spatial domain features and a frequency domain feature extraction module for extracting frequency domain features; and the spatial domain features and the frequency domain features are connected and input to a classification module to obtain a final prediction result. The method has the beneficial effects that the attention of the model to the irrelevant region is supervised and corrected through the mask, and a pixel-level label is used for guiding the model to face the synthesis region of the face, so that more accurate extraction of spatial features is ensured; the depth forgery detection method based on mask supervision has remarkable advantages in the aspects of detection accuracy, robustness, generalization ability, calculation efficiency, interpretability, overall performance optimization and the like, and a new solution is provided for the field of depth forgery detection.

The invention discloses a diffusion model self-guiding-based image or video generation method and system. The method comprises the steps that the input content is a noise potential variable and a future offset step length of a current time step; the diffusion model predicts the score network twice based on the input content to obtain a conditional score network of the current time step and the future offset time step; carrying out linear combination on the two conditional score networks to generate a combined conditional score network; denoising the noise potential variable of the current time step by using a combined condition score network; calculating and updating a potential variable of the previous time step according to a denoising result; in each time step, the self-guiding strategy is adopted to carry out denoising processing on the noise, and noise prediction is optimized; and after de-noising processing of all time steps, a de-noising potential variable is finally output as a final model generation result, and an image or a video is obtained. According to the invention, the quality of the generated image or video can be obviously improved.

The invention relates to the technical field of remote controller awakening, and discloses a remote controller awakening method, device and equipment and a storage medium, and the method comprises the steps: carrying out the multi-cascade coding processing of a key signal of a remote controller, obtaining the multi-cascade coding data, and generating a session key; performing adaptive carrier modulation to obtain a dynamic modulation signal; performing multi-band emission control processing and adaptive power control to obtain an emission control signal; performing low power consumption management on the remote controller system, creating a multi-stage dormancy strategy and a dynamic power consumption control scheme, and controlling the signal processor to receive the multiband optical signal; performing parallel processing and adaptive decoding on the multi-band optical signal to obtain decoded multi-cascade coded data; and based on the session key, performing multi-factor identity verification processing on the decoded multi-cascade coded data to obtain a secure wake-up instruction, thereby improving the reliability and security of wake-up, optimizing the power consumption management strategy, and enhancing the signal processing capability.

The invention relates to a network attack defense method and system for a distributed new energy grid-connected system, and the method comprises the steps: generating a dynamic defense strategy based on an IP address for the network abnormal flow and an attack sensing result when a new energy terminal is accessed; generating a random routing scheme based on a preset depth deterministic policy gradient; and performing dynamic defense in combination with the dynamic defense strategy and the random routing scheme. The method has the advantages that the comprehensiveness and comprehensiveness of network attack defense of the new energy grid-connected system are improved, and the network and data security of the new energy grid-connected system is guaranteed.

The invention provides a method and system for cleaning a ship oil flow pipeline, and the method comprises the steps: obtaining a first Reynolds number of flowing oil in each branch of the ship oil flow pipeline, and comparing the first Reynolds number with a critical Reynolds number; and acquiring the pollution grade of pollutants in each branch of the ship oil flow pipeline, and if the pollution grade is greater than a set value, judging that the ship oil flow pipeline needs series washing. According to the comparison result of the first Reynolds number and the critical Reynolds number, on the premise that series washing needs to be conducted on the ship oil flow pipeline, if the first Reynolds number is larger than or equal to the critical Reynolds number, flowing oil in all branches of the ship oil flow pipeline is in a turbulent flow state, the flowing oil can exert strong acting force on the pipe walls of all the branches during series washing, and therefore series washing is conducted on the ship oil flow pipeline. The inside of the pipeline can be cleaned more efficiently, and the serial washing efficiency is improved.

The invention provides a distributed power supply collaborative optimization power system reliability control method, which comprises the following steps of 1, obtaining real-time operation parameters of a power system, and obtaining an optimal output combination scheme of distributed power supply and conventional power supply collaborative optimization; 2, monitoring the running state of the distributed power supply in real time based on the Internet of Things technology; and step 3, scheduling and controlling active and reactive power output of the distributed power supply, and ensuring the power supply reliability of the important load. According to the technical scheme, important technical support is provided for constructing a power system optimization scheduling strategy with high robustness and high adaptability.

According to the program behavior intervention method based on the quantum state, the target program state is quantized, the quantum state is generated, synchronous monitoring of multiple program behaviors is achieved through the quantum entanglement effect, and the malicious behavior tendency of the program is detected by defining the quantum threshold function; the program quantum state is intervened by using a quantum control operator, so that the program state is adjusted and malicious behaviors are prevented; in addition, a random disturbance term is introduced to increase system complexity, and an intervention strategy is dynamically adjusted through multiple quantum state measurement and feedback, so that control parameters are optimized. According to the method, the problems of insufficient detection precision and weak defense capability when a traditional program behavior detection technology is used for processing complex program behaviors are effectively solved, the safety and the stability of the system are remarkably improved, and the detection and defense capability on complex malicious behaviors is enhanced.

The leg assembly is used for the humanoid robot, the leg assembly comprises a thigh shell, a first driving piece, a shank piece and a first connecting rod, and one end, in the first direction, of the thigh shell is used for being connected with a crotch assembly; at least part of the first driving piece is contained in the thigh shell; the shank part is rotationally connected with one end, far away from the crotch assembly, of the thigh shell in the first direction; the first connecting rod comprises a first end and a second end which are opposite, the first end is connected with the first driving piece, the second end is connected with the shank piece, and the first end is located on the outer side of the second end in the second direction; the first direction is the up-down direction of the humanoid robot, and the second direction is the left-right direction of the humanoid robot. According to the leg assembly, the size of the humanoid robot in the left-right direction can be reduced, the gravity center is not prone to shifting, and the humanoid effect is good.

The TOF module cover plate capable of avoiding glue overflow at least comprises a cover plate body, the cover plate body comprises a top wall and a side wall, the top wall and the side wall define a mounting cavity with an open bottom, the top wall is provided with a through slotted hole, and one side, close to the mounting cavity, of the top wall is provided with a plurality of glue grooves formed by sinking inwards at intervals in the edge direction of the slotted hole. A plurality of glue overflowing grooves formed by inwards sinking one side, close to the mounting cavity, of the top wall are formed between any glue groove and the groove hole at intervals, and a first retaining wall is formed between each glue overflowing groove and the corresponding glue groove. The glue box has the advantages that the structure is simple, the glue containing capacity is improved in a limited space, and glue overflowing caused by siphonage is avoided.

The invention discloses a multi-domain feature level fusion expression recognition method based on a graph neural network, and relates to the technical field of computer vision, and the method comprises the steps: firstly, carrying out the data enhancement of an original image, and carrying out the region division of a face according to three courts and five eyes based on key point detection; secondly, constructing a graph structure of the face, endowing each edge with a corresponding weight according to an area where the feature points are located, obtaining weighted distance information, and storing the weighted distance information in an adjacent matrix; then spectral domain features and spatial domain features are obtained through a graph neural network, and local features and context features are obtained through a multistage convolutional neural network and ViT; and then, multi-domain feature fusion is carried out through a feature fusion network based on structural feature guidance, and then an expression recognition result is obtained. Therefore, by adopting the multi-domain feature level fusion expression recognition method based on the graph neural network, the model can understand the change of the expression more deeply, the recognition capability is enhanced, and the robustness and the stability of the model are improved.

The invention discloses a dynamic S-box generation method and test system based on a spatiotemporal chaotic system, and the method comprises the steps: constructing an iteration matrix through employing a preset m sequence, and determining a coupling object index and a disturbance symbol of each grid in an iteration process; wherein each element of the iteration matrix represents a grid; performing dynamic disturbance on the initial chaotic sequence according to the coupling object index, the disturbance symbol, the index offset and the grid coordinate to obtain a dynamic chaotic sequence; initializing the two-dimensional coupling image lattice by using a dynamic chaos sequence, and carrying out iteration to obtain an iterated coupling image lattice; removing the previous N iteration results of the coupled image lattice after iteration; and performing numerical conversion on the chaotic sequence value obtained by the (N + 1) th iteration to construct an S box. According to the S box design and test method, the safety and the attack resistance of the S box are improved, and an efficient and reliable research tool can be provided for cryptography researchers.

The embodiment of the invention discloses an anti-quantum secure communication method, device and equipment applied to a public network channel. A specific embodiment of the method comprises the following steps: sending user authentication information to a communication receiving end; generating public and private key information of a sending end; based on public and private key information of a sending end, the following steps are executed: generating a sending end random number pair; sending the sending end public key information and a receiving end public key information request to a communication receiving end; performing public key scheme encryption on the random number pair of the sending end to obtain encrypted information of the sending end; generating master secret information; in response to determining that the current time meets the period time bar, based on the master secret information, executing a second loop step: determining a temporary session key and a temporary session effective duration; and in response to determining that the current communication duration is less than or equal to the valid duration of the temporary session key, performing encryption processing on the communication information. According to the embodiment, on the basis of resisting quantum attacks, the storage resource overhead of both communication parties can be reduced, and the communication encryption security can be improved.

The invention discloses a welding robot system based on visual guidance, which comprises an artificial potential field welding robot path planning algorithm, the artificial potential field welding robot path planning algorithm is combined with an ant colony algorithm, and the artificial potential field welding robot path planning algorithm and the ant colony algorithm comprise artificial potential field function simulation, ant colony algorithm simulation and path search, pruning strategy and path optimization mechanism. And the step operation system (GBOS) comprises three-dimensional model importing and welding seam positioning. According to the method, the welding path planning technology of the ant colony algorithm of the artificial potential field function is fused, the welding robot is guided to avoid obstacles by simulating attraction force and repulsive force, the optimal path can be accurately recognized, the coverage rate of the searched path to the environment can reach 74.63%, a good convergence curve can be obtained, and after the pruning strategy is introduced, the optimal path can be obtained. According to the method, the path can be optimized, redundant broken lines or path sections are removed, and optimization of the path length and complexity is achieved, so that the welding time and cost are reduced, and the welding efficiency is improved.

The invention discloses a network edge calculation method and a communication device, and relates to the technical field of network communication, and the method comprises the following calculation steps: inputting original data into a calculation processing module, and outputting processed data, encrypted data, a hash value of the encrypted data, an authentication code of a message and a verification result. And transmitting the encrypted data, the hash value of the encrypted data, the authentication code of the message and the verification result to a cloud server. The method has the advantages that the received data is changed firstly, so that the encryption efficiency and security of the data are improved, the data are encrypted in an encryption mode, the confidentiality of the data is guaranteed, the integrity verification of the data is carried out through the hash function and the message authentication code, and the security of the data is improved. According to the data transmission method and device, the data are ensured not to be tampered in the transmission process, all-directional protection of the data is achieved, meanwhile, the received data are verified in a verification mode so as to ensure the authenticity of the data, and dual guarantee of data transmission and verification is achieved.