83results about How to "Resolve distortion" patented technology

The invention discloses an intelligent driving environment sensing method and system based on multi-modal data fusion, and relates to the technical field of automatic driving, and the method comprises the steps: firstly obtaining the position information of a vehicle through a GPS, and starting an auxiliary positioning mechanism if the GPS signal is weak; multi-modal data acquisition and preprocessing: after an auxiliary positioning mechanism is started, a vehicle acquires surrounding environment information through an RGB camera, a laser radar and an IMU sensor; performing feature extraction on the RGB image by using an improved deep convolutional neural network, and generating a high-precision 3D bounding box through a region suggestion network; and multi-modal information is fused through a constructed efficient image reasoning framework. Based on multi-modal data fusion and an environment perception result, the method not only can carry out accurate target identification on the surrounding environment in the automatic driving process, but also can carry out global and local path planning.

The invention relates to a power data federated learning privacy protection method based on efficient homomorphic encryption. The method comprises the following steps: step 1, a client initializes a neural network model; step 2, the server side encrypts and updates local neural network model gradient parameters uploaded by the client side; 3, the server side executes gradient average aggregation operation, the server side does not need to decrypt, the encrypted gradient parameters uploaded by the client side are directly calculated, an average value is solved, and therefore global gradient parameters are updated; step 4, after each client receives the updated global gradient, carrying out decryption calculation operation; 5, loading the updated weight parameters of the neural network model into a local neural network model, and performing neural network model training by using a local data set; the problems that an existing encryption technology is large in operand, large in computing resource overhead and low in algorithm efficiency, and privacy leakage exists in the transmission and analysis process of power data can be solved.

The invention discloses a color adjustment method and device for a color electronic paper image, terminal equipment and a computer readable storage medium. The method comprises the steps of obtaining a to-be-processed display image of electronic paper; obtaining a color adjustment instruction; performing corresponding color adjustment on the to-be-processed display image according to the color adjustment instruction to obtain a pre-adjusted display image; a Floyd-Steinberg algorithm is adopted to carry out jitter processing on each pixel point in the pre-adjusted display image, and the updated pixel value of each pixel point is obtained; and for each pixel point, according to the updated pixel value and a preset color mapping table in the electronic paper, calculating a color distance to obtain a color distance set, and taking a color corresponding to a minimum value in the color distance set as a display color of the pixel point. By implementing the method and the device, the problems of limited color expression range and distortion when the electronic paper displays the image can be solved, and the visual experience of a user is improved.

The invention relates to the field of sensor communication, in particular to an intelligent sensor communication method. The method comprises the following steps: acquiring a physical quantity signal in an environment by using an intelligent sensor, generating a corresponding time domain signal, and converting the time domain signal into a frequency domain signal; based on the frequency domain signal, using a dynamic frequency hopping and power consumption perception optimization algorithm to select an optimal frequency; based on the optimal frequency, performing data block coding by using a nonlinear multi-level frequency coding algorithm; and evaluating the path state and the channel performance by using a path optimization selection algorithm, and calculating to obtain the optimal transmission path of each data block. The problems of frequency band overlapping and interference in sensor data acquisition are solved; the problems of fluctuation and inconsistency of sensor signals caused by changes of environmental conditions, reduction of signal reliability due to interference among different frequency channels, signal distortion due to frequency band interference among different data blocks, and fluctuation of channel quality and bandwidth of different paths along with time are solved.

The invention provides a wall interactive projection method and system based on CIM digital twinning, and is applied to the field of projection data processing. The environment change data of the wall projection area is collected in real time, the projection content can be dynamically adjusted in combination with the CIM model, the stability of the projection content in the physical space is ensured, meanwhile, the interaction content and the projection content input by the user are precisely synchronized by dividing the projection subareas and compensating the interaction influence in real time, and the user experience is improved. Projection distortion or jumping is automatically corrected under the environment change, accurate positioning and dynamic calibration are performed by using three-dimensional space coordinate data, and a projection path and angle are optimized, so that the distortion problem caused by angle change or shielding is avoided, and the stability of the projection effect and the user experience are improved.

The present invention belongs to the field of elevators, and specifically relates to a back-to-back guide rail installation tool for a backpack-type villa elevator, including a working platform substrate, a car guide rail installation assembly, and a counterweight guide rail installation assembly. The working platform substrate has at least one car guide rail installation position and at least one counterweight guide rail installation position. The car guide rail installation assembly is arranged on the car guide rail installation position, which is used to realize the detachable installation of the working platform substrate and the car guide rail and to calibrate the car guide rail. The counterweight guide rail installation assembly is arranged on the counterweight guide rail installation position, which is used to realize the detachable installation of the working platform substrate and the counterweight guide rail and to calibrate the counterweight guide rail. The present invention cleverly utilizes the relationship between the counterweight guide rail and the car guide rail to use a fixed and overall displacement method to calibrate the elevator guide rail. When the counterweight guide rail is calibrated, the car guide rail is naturally vertical, which reduces the installation workload and work intensity.

The invention relates to the technical field of power system fault detection, in particular to a factory power distribution system fault interval section determining method and system and terminal device.The factory power distribution system fault interval section determining method comprises the steps that when a power failure fault of a power distribution system is monitored, voltage and current data within a preset duration before the fault is obtained; comparing the phase current amplitude with a rated current threshold value, judging that the phase current amplitude is an external fault if the phase current amplitude does not exceed the limit, calculating a zero-sequence current if the phase current amplitude exceeds the limit, and comparing the zero-sequence current with a zero-sequence threshold value to distinguish a ground fault or an inter-phase fault; adaptively calling a phase difference analysis process according to an identification result, calculating a phase difference between a zero-sequence current and a zero-sequence voltage during a grounding fault, calculating a phase difference between a fault phase voltage and a fault phase current during an inter-phase fault, and judging whether a fault interval is internal or external according to a phase difference interval; and a structured message is automatically generated by integrating fault types, interval judgment and triggering criterion paths, so that quick and accurate fault positioning and handling are realized.

The invention discloses a tailoring device with a limiting structure for textile fabric processing, and relates to the technical field of textile fabric processing.The tailoring device comprises a main body, a limiting assembly, a power source, a processor, heat dissipation holes and an operation panel, the limiting assembly is connected with the processor through a signal line, and the operation panel is connected with the processor through a signal line; a limiting assembly is installed on the upper side of the outer wall of the body, a power source is installed on the lower side of the inner wall of the body, a processor is installed on the lower side of the inner wall of the body, heat dissipation holes are formed in the two sides of the outer wall of the body, and an operation panel is installed on the left side of the outer wall of the body. According to the textile fabric tailoring device, by installing the limiting assembly, the function of moving along with a textile fabric in the tailoring process of the textile fabric and limiting is achieved, the problem of distortion in the tailoring process of the textile fabric is solved, deformation in the tailoring process of the textile fabric can be prevented, the yield of tailoring processing of the textile fabric is increased, and the tailoring processing efficiency of the textile fabric is improved. And the subsequent manual processing cost is reduced, and the economic effect is improved.

The application discloses a double-reference parameter fusion DSQM non-cooperative signal separation method, device, equipment and medium, and relates to the technical field of digital signal processing.The method comprises the following steps: receiving an input signal and performing symbol extension; estimating an original input noise standard deviation, a weak signal and a signal-to-noise ratio, and calculating a right shift number and a dynamic weighted fusion coefficient; performing perturbation injection and phase compensation on the extended signal to generate an actual separation input signal, and estimating a noise standard deviation thereof; calculating an adaptive shift number k of the modified signal-to-noise ratio according to the original input weak signal and the actual separation input signal by using a dynamic weighted fusion algorithm; further calculating a left shift number m; amplifying the actual separation input signal by shifting the actual separation input signal left by m bits; sequentially performing right shift n-bit and left shift n-bit to extract a strong signal component; performing time domain subtraction to separate a weak signal component; respectively shifting m bits to the right to recover the magnitude; then respectively performing symmetric saturation truncation and asymmetric truncation to recover the bit width output. The application improves the signal separation quality.

The invention relates to the technical field of substrate glass kiln equipment, in particular to a large-extraction-amount kiln side breast wall supporting device and a mounting method thereof. The large-extraction-amount kiln side breast wall supporting device comprises a side breast wall supporting iron assembly and at least one supporting assembly, the side breast wall supporting iron assembly comprises at least two sections of side breast wall supporting iron, the side breast wall supporting iron is arranged in the length direction of a side breast wall, and a gap is formed between any two adjacent sections of side breast wall supporting iron; the supporting assembly is installed on a stand column in the kiln and used for supporting the side breast wall supporting iron assembly and fixedly connecting the side breast wall supporting iron assembly with the stand column in the kiln. According to the large-extraction-amount kiln side breast wall supporting device, the problem that an original integral supporting iron is distorted and deformed due to expansion in the process of heating up a kiln is solved, and therefore the safety of a kiln side breast wall and a crown in the whole process of heating up the kiln is guaranteed.

The present application relates to a kind of low-carbon park integrated energy system multi-agent collaborative carbon responsibility allocation method, steps include: the multi-source data of park is collected, with the minimum integrated energy system total operating cost as goal to carry out multi-energy flow optimization scheduling;With the dynamic correction model of two-way carbon flow considering the carbon flow transfer caused by energy storage charging and discharging process, dynamic carbon flow distribution matrix is constructed;Using load carbon responsibility decoupling allocation model, total load is decoupled into multiple sub-load categories, and different carbon responsibility allocation coefficients are given for each type of sub-load;With each interest party in park as the participant of game, Nash bargaining as collaborative optimization framework, maximum all participant utility gain as objective function, Shapley value as the basis for carbon responsibility allocation, multi-agent collaborative optimization is carried out, and the carbon responsibility allocation scheme of each subject is obtained in combination with user carbon responsibility allocation coefficient.Compared with prior art, the present application ensures the fairness, efficiency and operability of carbon responsibility allocation scheme.

The invention discloses a mine safety production monitoring method based on a multi-view optical system, and relates to the technical field of image processing, and the method comprises the steps: obtaining an original mine image through the multi-view optical system; obtaining internal parameters and external parameters of the multi-view optical system as parameter information of the multi-view optical system; acquiring a mine image by using the original mine image according to the parameter information of the multi-view optical system; and performing pipeline splicing processing according to the mine image to obtain a mine safety production monitoring result. The method not only solves the problems of camera distortion and monitoring dead angles in a traditional monitoring scheme, but also improves the quality and real-time performance of the monitored image through an intelligent image processing technology and NPU rectangular data processing, and has remarkable technical advantages and application prospects. And the safety and the working efficiency of mine production can be greatly improved by adopting a multi-view optical system.

The invention provides a water quality detection concentration calculation method and system, and relates to the field of water quality concentration detection.The method comprises the steps that after sample injection, quantification, preset reagent mixing, transferring and centrifugal color development are sequentially executed, photoelectric detection is continuously conducted on a target water sample for preset times, and the total phosphorus concentration, the total nitrogen concentration, the COD concentration and the ammonia nitrogen concentration for preset times are obtained; determining the actual total phosphorus concentration of the target water sample, and determining the actual ammonia nitrogen concentration of the target water sample; calculating an actual total nitrogen concentration matrix according to the initial total nitrogen concentration matrix and the total nitrogen implicit transfer matrix, and determining the actual total nitrogen concentration of the target water sample; and calculating an actual COD concentration matrix according to the initial COD concentration matrix and the COD implicit transfer matrix, and determining the actual COD concentration of the target water sample. According to the invention, a plurality of different index detections are integrated on the same chip, the accuracy of the detection results of the total nitrogen concentration, the COD concentration, the total phosphorus concentration and the ammonia nitrogen concentration is ensured, the detection time is shortened, and the detection efficiency is improved.

The invention discloses a screen segmentation method and a spherical screen, and relates to the field of special-shaped LED screens. In the prior art, a current mainstream LED dome screen splicing scheme is limited by pixel point distribution of small-spacing LEDs, zigzag arrangement is formed at the edge, obvious splicing gaps are generated after splicing, and especially at the top of a spherical surface, due to sharp contraction of a circular arc and denser pixel point distribution, an image is seriously deformed, and the splicing effect is poor. According to the technical scheme, the screen segmentation method is applied to the spherical screen and comprises the steps that a spherical surface where the spherical screen is located is segmented into six identical cambered surfaces; on each cambered surface, the cambered surface is cut along the longitude and the latitude according to the longitude and the latitude of the sphere where the cambered surface is located, and n * m small units are formed; the graphs where the edges of the small units are located comprise trapezoids with the edges being arc lines and triangles with the edges being arc lines. The method is suitable for being applied to design work of the small-spacing LED spherical screen.

The application discloses a plot quality evaluation and abnormality identification method and device based on multi-source forestry data fusion, relates to the technical field of forestry plot quality evaluation, and comprises the following steps: collecting multi-source forestry data and performing pretreatment, generating a standardized multi-source data set and performing dimension reduction processing, orthogonally decomposing the dimension-reduced data set to extract each principal component direction; projecting the standardized multi-source data set to each principal component direction to obtain each principal component score, performing denoising correction on the principal component score, and performing polarity correction on the corrected score; determining the weight of each principal component based on the polar-corrected principal component score and weighted calculating the comprehensive quality score of each plot; constructing the conditional probability distribution of the signal strength of each principal component and calculating the significance measure; determining whether each plot is abnormal according to the significance measure; and outputting diagnosis prompt information in combination with the comprehensive quality score and the abnormality determination result of each plot. Thus, the objectivity of plot quality comprehensive evaluation and the statistical credibility of abnormality determination are realized.

This application presents a method for providing clear, accurate, and intelligent AI representations of complex, multi-element images. It employs an improved Faster R-CNN model as the image processing model, replaces VGG-16 with DenseNet for feature extraction, optimizes bounding box selection through dimensional clustering, and enhances model stability by predicting relative grid cell coordinates. A text generation architecture is constructed based on a hierarchical recursive network. The first layer, a global long short memory layer, determines the topic vector based on the fusion vector to increase the number of descriptive sentences. The second layer, a sentence BRNN layer, generates accurate and rich descriptive sentences. A fully convolutional localization layer effectively integrates and transfers block features, constructing an end-to-end architecture that collaboratively optimizes image processing and text generation. This method addresses the problems of high model complexity, inaccurate bounding box localization, and insufficient association of descriptive sentences in existing technologies, generating denser, more accurate, and natural descriptive language that conforms to language habits. It is suitable for various application scenarios such as visual intelligent question answering and video subtitle generation.

The application belongs to the technical field of point cloud processing, and discloses a multi-modal semantic mapping method and system based on a master-slave architecture. The method comprises the following steps: S1, extracting feature points in point cloud data and calculating the covariance matrix of the neighborhood point cloud of each feature point; S2, solving the optimal pose corresponding to the minimum of the target function; S3, converting the obtained point cloud data into a world coordinate system using the optimal pose, fusing all frame point cloud data in the world coordinate system to form a three-dimensional point cloud map, projecting the three-dimensional point cloud map to a camera coordinate system, and inputting the three-dimensional point cloud map projected to the camera coordinate system into a semantic segmentation neural network to obtain the semantic label corresponding to each point cloud in the three-dimensional point cloud map, thereby achieving semantic mapping. Through the application, the problems of weak dynamic adaptability, insufficient model generalization ability and long-term running precision decay of the traditional three-dimensional semantic mapping method in a dynamic complex scene are solved.

The application provides a dynamic modeling method and system for multi-field skill AI automatic semantic labeling, and relates to information retrieval, medical health and cross-field skill fusion scene. Through nonlinear dynamic modeling and biomedical stability control, the semantic label automatic generation and real-time matching of talent skill and health intervention effectiveness are realized. The application innovatively integrates the sound wave frequency of fetal brain promoting music method, maternal action instruction and blood type nutrition scheme into the "skill-time" frequency domain space, combines the Gevrey smoothing operator and fractional derivative evolution equation, accurately separates the temporary fetal movement fluctuation and low-frequency core characteristics, and breaks through the characterization limitation of traditional models on nonlinear biological coupling effect. By introducing the "intervention effect-time" frequency domain component and local dependency constraint, the short-term effectiveness and long-term value of music intervention on fetal neurodevelopment are quantified, and the model distortion problem caused by extreme data or cross-field parameter resonance in traditional methods is solved.

The application provides an image fusion method and device and a storage medium, and belongs to the technical field of image fusion.The method comprises the following steps: importing a high-resolution hyperspectral image to be fused, a high-resolution multispectral image to be fused, a plurality of original high-resolution hyperspectral images and an original high-resolution multispectral image; performing down-sampling processing on the original high-resolution hyperspectral images respectively to obtain original low-resolution hyperspectral images; and constructing a training model, wherein the training model comprises a feature analysis network, a target vector field construction network and a predicted vector field construction network.The application effectively avoids the spectral distortion problem caused by the difference in spectral response of a sensor in a traditional fusion method, and overcomes the defects of lack of constraints in the intermediate process and insufficient stability of results in a single-step regression method.

The invention relates to the technical field of data processing, in particular to a BIM model-based multi-party cooperative construction management and control method and system, and the method comprises the steps: obtaining the point cloud data of a real-time construction project, and determining the optimal thick side length of a voxel block according to the down-sampling rate of the point cloud data and the unit side length of the spatial distribution of the point cloud data; voxelization is carried out on the point cloud data based on the optimal coarse side length to obtain a plurality of voxel blocks, and the average density of point cloud distribution of the point cloud data in the voxel blocks is determined; determining the optimal side length of the voxel block according to the average density, the down-sampling rate, the optimal coarse side length and the unit side length, and performing down-sampling on the point cloud data based on the optimal side length to obtain a down-sampling project point cloud set of the real-time construction project; and constructing a BIM model of the real-time construction project by using the down-sampling project point cloud set, and managing and controlling the real-time construction project by using the BIM model. Therefore, the construction accuracy of the BIM three-dimensional model is improved.

The application discloses a kind of blood sampling test tube outer wall laser printing information and bar code device and its using method in mechanical technical field, including base and test tube, and test tube is prefabricated with engraving layer, and test tube holder and laser engraver are arranged on base;Test tube holder includes electric screw, and electric screw is slidably connected with mounting platform, and mounting platform is provided with cylinder, and the output shaft of cylinder is fixedly connected with mounting plate, and mounting plate bottom is rotatably connected with rotating rod, and rotating rod is provided with driver for driving rotating rod to rotate, and rotating rod bottom is fixedly connected with mechanical gripper for grabbing test tube;Electric hatch is opened in base, and material delivery bin for storing test tubes of various colors and controller are arranged in base.The technical scheme of the application can perform laser printing on test tube, and improve label marking efficiency.

The application provides a curved panel defect detection method and system. The method comprises the following steps: acquiring multiple perspective images of a curved panel from multiple perspectives; preprocessing each perspective image, performing projection transformation based on the preprocessed perspective image, and reconstructing a panoramic development map of the curved panel; performing defect recognition and detection based on the panoramic development map; and outputting a defect recognition and detection result of the curved panel. The method can accurately locate and determine defects, and reduces the false detection rate and the missed detection rate.

This application provides a method, equipment, medium, and product for testing power line signals, relating to the field of signal processing technology. The method includes: dividing the power frequency cycle into N test time cycles using a voltage zero-crossing detection unit and generating corresponding N wideband test data packets; sequentially sending the wideband test data packets within M consecutive power frequency cycles to obtain M test signals, which are then collected to form N sub-signal sets; calculating the fluctuation variance of each sub-signal set and generating a confidence coefficient; synthesizing N reference sub-signals and dividing them into high-confidence sub-signals and sub-signals to be repaired based on the confidence coefficients; reconstructing and repairing the sub-signals to be repaired using the channel response characteristics of the high-confidence sub-signals to generate repaired sub-signals; and finally concatenating the high-confidence sub-signals and repaired sub-signals in time sequence to obtain the final test signal. This solution solves the technical problem of how to obtain high-precision power line test signals throughout the entire power frequency cycle.

The invention relates to the technical field of reactors, in particular to a corrosion product deposition growth prediction method, a storage medium and electronic equipment. The corrosion product deposition growth prediction method comprises the following steps: establishing a three-dimensional model, and carrying out mesh generation on the three-dimensional model to obtain a mesh model; obtaining the thickness increment of the corrosion product; converting the thickness increment of the corrosion product on the grid unit surface into a node movement amount by using a node movement algorithm, and driving grid nodes to displace in each time step length according to the calculated node movement amount based on a dynamic grid technology; obtaining an updated grid model; the deposition rate and the erosion rate of the corrosion product are solved on the updated grid model, and the thickness increment of the updated corrosion product is obtained. The application of the invention can improve the safety of the reactor, and has important values for protecting the integrity of the fuel rod, preventing the leakage of radioactive substances and ensuring the safe operation of the reactor.

The invention discloses a power transmission and transformation equipment component mechanism and data mapping model construction method, which comprises the following steps of: 1, establishing a model simplification rule, and simplifying a power transmission and transformation equipment component geometric model; 2, an electrostatic field module, a solid heat transfer module, a solid mechanical module and a physical field coupling module are built based on a finite element analysis platform, and accurate simulation of the multi-physical field coupling process is achieved through parameter setting and data interaction between the modules; step 3, establishing a one-to-one correspondence relationship between simulation data and physical mechanisms by taking the core physical mechanism of the power transmission and transformation equipment element as a link, and forming a structured mechanism-data mapping system; 4, performing mapping verification and iterative optimization to form a closed-loop optimization mechanism; according to the method, calculation precision and analysis efficiency are both considered, a multi-physics field coupling mechanism is integrated, and an accurate power transmission and transformation equipment element mechanism and data mapping model is constructed.

The present invention discloses a method for detecting multiple types of defects on the inner wall of a corrugated pipe based on deep learning, comprising the following steps: S1, collecting an initial image of the inner wall of the corrugated pipe and images of the inner wall of the corrugated pipe under different lighting environments to generate an original image; S2, preprocessing the original image to generate a preprocessed image; S3, extracting the overall surface geometric features of the inner wall of the corrugated pipe from the preprocessed image through a pre-trained ResNet101 network; S4, inputting the preprocessed image and the overall surface geometric features into an improved StyleGAN3 network to generate an enhanced image; S5, inputting the enhanced image into a multi-scale adaptive deep network to output a defect detection result; S6, post-processing the defect detection result to output a final defect detection result. The present invention adopts deep learning and geometric feature embedding technology to realize accurate detection of defects on the inner wall of the corrugated pipe, and has the advantages of high precision, strong robustness and comprehensive adaptability.

The invention provides a traffic state prediction method and device, a terminal and a medium, and relates to the technical field of intelligent traffic, and the method comprises the steps: recognizing a key traffic road section as a key node based on the historical traffic state data of each traffic road section; constructing a double-layer traffic network structure according to an identification result; a first traffic network structure in the structure comprises all traffic road sections and a spatial connection relation thereof, and a second traffic network structure only comprises key traffic road sections serving as key nodes and a spatial connection relation thereof; and extracting network-level spatial features and key node spatial features based on a double-layer traffic network structure, guiding fusion of the network-level spatial features and the key node spatial features based on prior knowledge information related to each traffic road section, and predicting the traffic state based on the fused spatial features. According to the method, the key node characteristics in the traffic network can be combined, the traffic road segment priori knowledge is utilized to guide feature fusion, and the traffic state prediction accuracy and stability are improved.

This invention belongs to the field of image processing technology, specifically relating to a method for evaluating the lifespan of drilling tools based on image fusion. The method includes acquiring surface images, infrared images, and laser point cloud data during the operation of a mud pump; extracting low-frequency and high-frequency components from the surface and infrared images, and performing low-frequency fusion and high-frequency fusion respectively to obtain a fusion result; determining the actual wear depth based on the laser point cloud data, and correcting the fusion result based on the actual wear depth to obtain a corrected fusion feature; constructing a wear quantization model using the laser point cloud data, and determining wear quantization parameters based on the wear quantization model; inputting the corrected fusion feature and wear quantization parameters into a lifespan prediction model to output the corresponding remaining lifespan. According to the solution of this invention, the distortion problem of two-dimensional images is effectively solved, the influence of viewing angle on two-dimensional images is reduced, and the accuracy of mud pump lifespan prediction is improved.

The present application relates to the field of intelligent prediction, and particularly relates to a weak adhesion hydrogel adhesion performance prediction method and system, comprising the following steps: S1: obtaining hydrogel material parameters and interface state information and preprocessing; S2: identifying state labels through a classification model to form an extended data set with state labels; S3: constructing a segmented physical model system driven by the interface state to obtain state-adaptive physical prediction results; S4: embedding the state-adaptive physical prediction results into a physically-constrained neural network structure to construct a prediction model and obtain model prediction results; S5: using a multi-model collaborative fusion mechanism, comprehensively considering the state-driven physical model output and the model prediction results, and obtaining the final prediction value through a weighted fusion manner; S6: taking the final prediction results and model internal parameters as inputs, performing feature contribution degree analysis, identifying key influencing factors, and outputting parameter optimization suggestions. The present application realizes high-precision and high-generalization-capability prediction.

The invention discloses an emergency monitoring robot combined with Raspberry Pi distortion removal monitoring, and relates to the technical field of image processing, the emergency monitoring robot comprises a visual identification module, and the visual identification module is used for carrying out distortion removal processing on an image collected by the emergency monitoring robot and identifying a target object in the image; the visual identification module comprises an image acquisition unit, a distortion removal processing unit and a target identification unit, the image acquisition unit is connected with the distortion removal processing unit through a network signal, and the distortion removal processing unit is connected with the target identification unit through a network signal. According to the invention, the visual identification module is designed, so that the function of identifying the target object in the image is realized, the problem of image distortion caused by camera distortion is solved, the availability and accuracy of the image are improved, and the identification precision is improved.