145 results about "Laser data" patented technology

The invention discloses a high-precision laser positioning unpacking path planning system, and belongs to the technical field of robot automatic control and industrial automation. The system comprises a data synchronization module used for multi-sensor hardware synchronization and data alignment; the high-precision positioning module is used for outputting a precise pose based on environment skeleton characteristics and sliding window optimization; the semantic map construction module is used for fusing vision and laser data to generate a dynamic semantic grid map; the global path planning module is used for planning a smooth path in the skeleton channel by utilizing a mixed potential field improved A * algorithm; the motion control module is used for realizing closed-loop motion control and safety monitoring through model prediction and tracking; and the operation execution module is used for finishing millimeter-level precise stopping of an operation point by adopting visual servo and triggering unpacking operation. According to the method, the positioning robustness under dynamic shielding is improved through the environmental skeleton features, the safety and the high efficiency of the path are ensured by utilizing semantic understanding and intelligent planning, and the full-process automation from navigation to precise operation is realized.

The invention discloses a weld joint three-dimensional structure detection device based on vortex laser, and relates to the field of laser welding, and the weld joint three-dimensional structure detection device comprises a vortex laser generation system used for generating and emitting vortex laser; the image acquisition unit is used for collecting returned laser data projected to the welding seam and carrying out dynamic scanning and data acquisition; the motion control and feedback system is used for adjusting the relative position of the laser welding head and the workpiece in real time and dynamically adjusting parameters such as welding power and scanning speed according to the collected data result; and the data processing and analyzing platform is used for carrying out image processing and feature extraction on the collected multi-dimensional data and modeling the morphology of the welding sample. The vortex laser serves as a light source for welding seam detection, high-precision, high-depth, multi-parameter and strong-interference-resistance three-dimensional welding seam modeling is achieved, the problems that in the long and thick plate welding process, welding seam detection parameters are few, precision is low, and real-time adjustment cannot be achieved are solved, and meanwhile the device has the performance characteristics of being high in detection precision, high in efficiency and comprehensive in structural parameter.

The invention relates to an automatic guided vehicle (AGV) adaptive path tracking method based on MPC. The method comprises the following steps: S1, establishing a kinematic discretization error model based on the kinematic characteristics of the two-wheel differential AGV, processing a continuous kinematic equation by adopting an Euler discretization method, expressing a dynamic change relationship between a transverse deviation distance and an angle deviation in a state-space equation form, and generating a kinematic discrete state-space model; the method has the advantages that the continuous equation is processed by establishing the kinematics discretization error model and adopting the Euler discretization method, the deviation dynamic relation is expressed through the state space, the calculation process is simplified, the precision is kept, sensor data are fused, the wheel type odometer, IMU and laser data are integrated through the extended Kalman filtering algorithm, and the precision is improved. The real-time position and angle deviation are calculated, the positioning accuracy is improved, a model prediction controller objective function is designed, a weight matrix and boundary constraint are introduced according to constraint conditions, and the effect of obstacle avoidance constraint is combined.

The invention belongs to the technical field of lake remote sensing monitoring, discloses a lake water level and water reserve dynamic monitoring method based on satellite-borne active and passive remote sensing information fusion, and effectively solves the problem that lake information in remote areas is difficult to efficiently and accurately obtain. Comprising the following steps: S1, preprocessing laser data; s2, preprocessing the multispectral data; s3, taking the ICESat-2 data as a core, combining with a multispectral image covering a corresponding region, and constructing a lake bottom DEM inversion data set based on neighborhood features; s4, constructing and training an elevation inversion convolutional neural network model; s5, predicting and generating a lake bottom DEM by using the trained elevation inversion convolutional neural network model; and S6, calculating the absolute water volume of the lake according to the DEM at the bottom of the lake, and carrying out lake water level-water reserve monitoring. According to the invention, long-time-sequence hydrological observation of remote region lakes can be effectively realized with low cost.

The invention discloses an underground mine positioning and mapping trajectory optimization method based on underground control point constraint, and belongs to the field of positioning and mapping, and the method comprises the steps: respectively obtaining GNSS data, laser data, IMU data and image data; according to the GNSS data, the laser data, the IMU data and the image data, completing initialization of a laser radar-vision-inertial odometer system and GNSS initialization; extracting point cloud features of the distributed underground control points in real time based on an RANSAC algorithm, and generating a global coordinate constraint factor of nonlinear least square optimization; and performing relative attitude calculation according to the global coordinate constraint factor to complete underground mine positioning and mapping trajectory optimization. According to the method, the problems of accumulated error overrun and point cloud dislocation layering easily caused by pose drift in a special environment of an underground mine are solved.

The invention relates to the technical field of laser data transmission, and provides a multi-dispersion laser system and a control method thereof. The control method of the multi-dispersion laser system comprises the following steps: establishing a first model and a second model; in the initial state, the pumping power of the pumping light source is adjusted, and a first target pulse interval and first target power are determined in combination with the first attenuation coefficient, the second attenuation coefficient and the first model; determining a second target power, a second target pulse interval and a target attenuation coefficient by a second model in combination with the fact that the second attenuation coefficient is zero; the real-time pumping power is controlled to be first target power, the first real-time attenuation coefficient is zero, the second real-time attenuation coefficient is zero, the first laser is output in a first state, and the second laser is output in a third state; the third state comprises a soliton dissipation state; the real-time pumping power is controlled to be second target power, the first real-time attenuation coefficient is the target attenuation coefficient, the second real-time attenuation coefficient is zero, the first laser is output in the first state, and the second laser is output in the second state.

The invention provides an indoor robot target detection method and system based on multi-modal tracking and a medium, and belongs to the field of robotics.The method comprises the steps that laser data processing is optimized through static furniture learning and filtering and dynamic region clustering, and the visual anti-illumination interference capacity is combined; a dynamic cost matrix matching strategy based on target motion characteristics is adopted to obtain a stable trajectory of a target person; kalman filtering trajectory prediction and target appearance feature matching are fused, and it is ensured that tracking can be rapidly recovered after the target is temporarily lost; a human-guided following mapping mode is adopted, a moving track of a tracked target is used as dynamic guidance, dynamic and static point clouds are separated, and a high-precision environment map with semantic tags is constructed; according to the invention, a following state monitoring and active service linkage mechanism is established, scene decision is carried out by identifying a target attitude and combining a semantic map, corresponding health monitoring and safety services are triggered, and the crossing from passive tracking to active intelligent services is realized.

The invention discloses an online quality detection method and system for a stranded cable. The system comprises a data acquisition module, a data processing module and a quality detection module, relates to the technical field of cable quality detection, and solves the technical problem that the accuracy of cable quality detection in the wire drawing and stranding process is not high due to the fact that the acquisition frequency of a sensor is dynamically adjusted according to the movement speed of a stranded cable in the prior art. According to the invention, the movement speed of the wire rod is obtained in real time and the detection frequency is dynamically adjusted, so that the synchronous matching of the detection process and the production rhythm is ensured, and the problem of missing detection or repeated acquisition caused by traditional fixed frequency detection is avoided; the cable stranded wire is divided into a plurality of detection sections, and multi-dimensional data acquisition and analysis are performed in combination with an image recognition algorithm and a laser detection technology, so that the defect recognition precision and the detection efficiency are improved; and furthermore, the quality condition is comprehensively judged based on the defect area and the laser data, so that the accuracy of quality evaluation of the cable stranded wire is improved.

The invention discloses an enamelled rectangular wire online detection system and method, relates to the technical field of enamelled rectangular wire quality inspection, solves the problems of limited range and poor precision of a current enamelled wire detection mode, and comprises a laser data acquisition unit, a laser initial inspection unit, an image acquisition unit, an image monitoring unit and a defect identification module, the laser initial detection unit is used for detecting and analyzing external monitoring data of the enameled rectangular wire, abnormal laser monitoring data, obtained through analysis, of the enameled rectangular wire are sent to the database module, and the image monitoring unit is used for detecting and analyzing surface image data of the enameled rectangular wire, obtaining abnormal image monitoring data of the enameled rectangular wire and sending the abnormal laser monitoring data to the database module. The defect identification module is used for analyzing the defect condition of the enameled rectangular wire in combination with the abnormal laser monitoring data and the abnormal image monitoring data, the abnormal degree of the enameled rectangular wire is obtained through analysis, and comprehensive detection and accurate detection of the enameled rectangular wire are achieved.

The invention relates to a sensor cleaning control method of an unmanned vehicle and a related device thereof, a sensor of the unmanned vehicle is provided with a pollution monitoring unit, and the method comprises the following steps: obtaining multi-element pollution data collected by the pollution monitoring unit; wherein the pollution detection unit at least comprises an image acquisition module and a laser detection module; the multi-element pollution data comprises image data acquired by the image acquisition module and laser data acquired by the laser detection module; the image data and the laser data are analyzed to obtain a pollution result of the sensor, and the pollution result is used for indicating the pollution degree of the sensor; when the pollution result indicates that the sensor is in a pollution state, determining a cleaning strategy based on the multi-element pollution data; and controlling a cleaning system of the unmanned vehicle to clean the sensor according to the cleaning strategy. According to the scheme provided by the invention, the pollution type and degree of the sensor can be accurately identified, a targeted cleaning strategy is adopted, and the cleaning efficiency and the cleaning capability are improved.

The invention belongs to the field of robot path planning, and particularly relates to a method and system for dynamically covering a cleaning path of a cleaning robot, and the method comprises the following steps: sending a cleaning region task instruction to a task processing module; the task processing module controls the cleaning robot to sequentially execute cleaning tasks according to the task execution sequence; the motion control module sends corresponding cleaning path planning commands to the path planning module according to different cleaning task types; the sensing module processes the laser data and the camera data and then sends the laser data and the camera data to the path planning module; the path planning module updates the environmental information of the cleaning area, plans a cleaning path in real time based on the updated environmental information of the cleaning area, and sends the planned path to the motion control module; and the motion control module controls hardware of the cleaning robot to dynamically and intelligently execute cleaning operation. A dynamic path planning technology is used, so that the cleaning robot can dynamically plan the cleaning path in real time in a dynamic cleaning area, and the cleaning operation is efficiently completed.

The invention relates to a line laser contourgraph data consistency evaluation method and device which are applied to the field of optical precision measurement, and the method comprises the steps: obtaining point cloud data generated when a line laser contourgraph scans a measured plane of a measured target object; determining a spatial relationship between the measured plane and an ideal horizontal plane based on the point cloud data; determining an ideal horizontal plane based on an optical axis formed by a line laser profiler; constructing a rotation matrix between the original coordinate system of the point cloud data and the coordinate system of the ideal horizontal plane according to the spatial relationship between the measured plane and the ideal horizontal plane; correcting the point coordinate of the original laser line in the point cloud data through the rotation matrix to obtain the corrected point coordinate of the laser line; and evaluating the data consistency of the line laser contourgraph based on the corrected point coordinates of the laser lines. According to the invention, the problem that the accuracy of the consistency evaluation of the overall line laser data measured by the line laser contourgraph needs to be improved is solved.

The invention discloses a shield tunneling machine attitude deviation measurement method based on laser-vision deep learning heterogeneous fusion, and the method comprises the steps: employing double sensors of a laser target and a vision camera, carrying out the deployment of the sensors, and collecting data, which comprises laser data and vision data; preprocessing the data, wherein the preprocessing comprises laser data preprocessing and visual data preprocessing; the laser data preprocessing comprises laser data coordinate conversion, laser target attitude deviation calculation and noise filtering; the visual data preprocessing comprises image preprocessing, mark point detection, camera calibration and world coordinate calculation, and visual measurement attitude deviation calculation; a deep learning heterogeneous fusion model is constructed, the deep learning heterogeneous fusion model comprises a laser absolute feature branch, a visual spatial-temporal feature branch, an attention fusion layer and an output layer, and the model is trained; and inputting the preprocessed real-time data into the trained deep learning heterogeneous fusion model, and carrying out dynamic calibration on an output result.

The invention belongs to the technical field of welding robots, and particularly relates to a welding robot error correction method based on vision and laser data fusion. The method comprises the following steps: firstly, calibrating parameters on a dual-focus conjugate plane through a binocular camera, establishing a three-dimensional distortion correction model, constructing depth information in combination with a binocular triangle principle, completing laser light plane parameter calibration through an SVD (Singular Value Decomposition) method, and outputting large-field-of-view and high-depth-of-field laser-assisted vision measurement data; secondly, designing a three-dimensional target containing sphere characteristics, initially calibrating hand-eye parameters based on a fixed-point variable attitude method, establishing a hand-eye relationship and distance error identification model, and solving errors to realize robot-vision integrated online calibration; and finally, extracting a welding seam center point in a layered manner, planning a track through unit quaternion interpolation and a mixed curve, and writing an interface to output a welding control scheme. The problems that a traditional method is low in calibration precision and difficult in light bar extraction are solved, the measurement and positioning precision of the welding robot is greatly improved, and the high-precision welding requirement of a large component is met.

The invention relates to the technical field of unmanned aerial vehicle positioning, discloses an unmanned aerial vehicle positioning method and device fusing visual data and laser data, electronic equipment, a computer readable storage medium and a computer program product, and is used for solving the problems that an unmanned aerial vehicle positioning method in the prior art is greatly influenced by scene environment feature changes, and the positioning accuracy is high. And universality and robustness are not good. The method comprises the following steps: acquiring visual data and laser data, and generating a visual point cloud and a laser point cloud; respectively extracting feature information of the visual point cloud and the laser point cloud, and matching and aligning the visual point cloud and the laser point cloud according to the feature information; evaluating position measurement errors of the visual point cloud and the laser point cloud in real time, and dynamically calculating weight parameters; and carrying out weighted fusion positioning in real time based on the visual image and the laser image in combination with the weight parameter to obtain unmanned aerial vehicle positioning information. According to the method, the unmanned aerial vehicle can adaptively fuse visual and laser data according to an actual scene for accurate positioning.

The application provides a laser curtain wall image optimization system and method based on laser ranging analysis, and relates to the technical field of laser curtain walls.The application generates snowflake point probability and non-snowflake point probability by performing probability analysis on a snowflake point set and a non-snowflake point set, improves precision through variance analysis, and introduces a variance analysis module for snowflake points and non-snowflake points, which helps to deeply understand the intensity distribution characteristics of different reflection sources, outputs a reflection intensity threshold based on the size of the variance, optimizes the precision of laser ranging, and effectively identifies laser points that have passed through snowflake reflection and laser points that have not passed through snowflake reflection through the setting of a laser point cloud collection module and a division module, thereby realizing accurate classification of laser data; considering the possible jitter of the laser equipment under the influence of wind, the jitter data of the equipment are acquired in real time through a jitter collection module and an analysis module, and correlation analysis is performed with the reflection intensity threshold, thereby ensuring the authenticity of the data.

The invention belongs to the field of robot control, and provides a robot path planning method and system based on a target attention point, and the method comprises the steps: carrying out the environment field modeling based on a Gaussian process according to laser data obtained by a robot; based on a leading edge detection and clustering algorithm, generating target attention candidate points at known-unknown boundaries of an environment field, and taking prediction variances at the candidate points as measurement of field uncertainty; constructing a utility function fusing map information gain, field information gain and path cost, and selecting a target attention point from the candidate points by maximizing the utility function; and generating an optimal path by using a TEB path planner. According to the method, autonomous sampling of the robot can be effectively guided, information acquisition is maximized while the advancing path is optimized, and the precision and efficiency of environment field reconstruction are remarkably improved.

This application discloses a semantic map construction method, apparatus, electronic device, and storage medium. The method includes: acquiring source data for the semantic map, including image data, laser data, inertial navigation data, and satellite positioning data; determining the vehicle pose corresponding to the image data based on the image data and inertial navigation data, and determining the vehicle pose corresponding to the laser data based on the laser data and inertial navigation data; constructing a first nonlinear optimization model based on the vehicle pose corresponding to the image data, the vehicle pose corresponding to the laser data, and the satellite positioning data to determine the optimized vehicle pose; determining the three-dimensional semantic information corresponding to the image data based on the image data and laser data; and constructing a semantic map based on the optimized vehicle pose and the three-dimensional semantic information. This application optimizes the vehicle pose by combining data from multiple sensors, improving the accuracy of semantic map construction, and improves the efficiency of semantic map construction by combining image data and laser data.

The invention relates to the technical field of charging positioning, and discloses a marine automatic charging data communication system based on optical positioning, which comprises a region division unit, an optical positioning unit, an offset calculation unit, a correction unit, a data fusion unit and a docking unit. The problem of automatic charging of the unmanned ship in a dense fog environment is solved, the visibility attenuation coefficient is calculated through laser data, the laser power is adjusted as required, stable transmission of signals in dense fog is ensured, closed-loop cooperation from optical positioning to a docking unit is ensured, and the whole charging process is controlled through data such as positioning coordinates and attitude offset. According to the scheme, the influence of water surface fluctuation can be accurately counteracted, safe area setting and contact force dynamic adjustment during trajectory planning are achieved, hard butt joint is avoided, the loss of charging equipment is reduced, and compared with visual recognition depending on light, the scheme can still achieve accurate butt joint under low visibility, and the accuracy and stability of automatic charging are improved.

The invention discloses a display system based on laser point cloud calibration, the display system comprises a display interface and a controller, and the display interface comprises a point cloud display area and a statistical display area; the controller is used for executing the following steps: step 1, controlling the laser data to be displayed as a real-time point set in the point cloud display area; a first algorithm module controls a target frame to be displayed in a point cloud display area and frames a to-be-calibrated point set; 2, processing the to-be-calibrated point set into a template point set by a second algorithm module, displaying the template point set in a point cloud display area, and removing miscellaneous points from the template point set in an interface interaction mode to obtain an optimized template point set; 3, a third algorithm module controls the optimization template point set to be matched with the real-time point set, and poses of matching points are generated and displayed in a statistical display area; and 4, calculating a statistical evaluation score corresponding to the matching point by a fourth algorithm module, evaluating the stability of the laser point cloud, and displaying the statistical evaluation score in a statistical display area.

The invention discloses a laser detection device capable of continuously measuring the wall thickness of a pipe, and relates to the technical field of laser measurement, the laser detection device comprises a measurement frame, the inner wall of the measurement frame is slidably connected with two mechanical clamping rods, a laser measurement plate and a brush plate are arranged above the measurement frame, and the inner wall of the measurement frame is also provided with a data storage element. According to the device, firstly, through the arrangement of the brush plate, circumferential cleaning work can be conducted on a pipe end opening, chipping attachments generated by cutting on the pipe end opening are removed, the accuracy of follow-up data collection of the laser measuring plate is conveniently maintained, meanwhile, the brush plate has the characteristic of moving in advance, and the practicability of the device is improved. The pipe wall thickness data collected by the laser measuring plate on a detection path can be ensured to be accurate data after cleaning, the brush plate and the laser measuring plate can keep a certain distance, the brush plate and the laser measuring plate do not interfere with each other, and in the rotating process of the laser measuring plate, the phenomena of chipping falling and the like do not occur, so that the pipe wall thickness can be accurately measured. And the accuracy of laser data recording is influenced.

The present invention discloses a method, device, equipment and storage medium for monitoring ice thickness on power transmission lines. Multiple optical fibers are laid along the extension direction of the power transmission line. A fiber grating sensor is provided in the optical fiber at the connection between the power transmission line and the tower. A wind speed sensor is also provided on the tower. The method includes: obtaining laser data signals reflected by each fiber grating sensor uploaded by a laser receiver, and determining the wavelength of the laser reflected by the fiber grating sensor based on the laser data signal. Based on the wavelength of the laser reflected by each fiber grating sensor, the actual stress of the power transmission line between two adjacent towers is calculated. A stress function of the stress of the power transmission line between two adjacent towers with respect to ice thickness is constructed. The actual stress is substituted into the stress function to obtain the ice thickness. The present invention utilizes fiber grating sensors to achieve high-precision real-time monitoring of the icing condition of power transmission lines, which is particularly suitable for extreme weather conditions and greatly improves the accuracy and real-time performance of monitoring data.

The application discloses a kind of multi-source asynchronous laser scanning data fusion and three-dimensional imaging method and system based on heterogeneous embedded system, the method is for the time dislocation between high-frequency laser data and low-frequency position feedback, through Kalman filtering algorithm, realize the motion state smooth redrawing of microsecond level;Using the hardware parallelism of FPGA, the PL end synchronous sampling ADC data and encoder position are implemented nanosecond level timing alignment;In three-dimensional space projection stage, introduce adaptive weight fusion based on space topological relation, solved the artifact distortion problem when point cloud is projected to grid;DMA interrupt driven ping-pong buffer mechanism ensures that ADC data acquisition and PS end processing are full-pipeline parallel, system throughput is not affected by CPU processing delay.The application effectively solves the data fusion distortion problem caused by heterogeneous multi-source data sampling rate mismatch and timing asynchrony, improves the continuity, real-time and spatial resolution of laser scanning imaging.

The invention provides a simulation environment construction method and device, electronic equipment and a storage medium. The method comprises the following steps: determining a map point cloud based on a grid map; constructing an obstacle point cloud and a virtual mobile device in the map point cloud; and within a preset range of the virtual mobile device, determining laser data of the virtual mobile device based on the map point cloud and the obstacle point cloud. According to the method, the simulation test efficiency and the utilization rate of simulation resources can be improved.

The present application provides a method, device, equipment and storage medium for identifying a long corridor environment. The recognition system first collects laser data and calculates the coordinate mean of all first laser points in the neighborhood corresponding to each laser point in the current frame laser data. Then, the covariance matrix corresponding to each laser point is constructed according to the coordinate mean of each neighborhood, and the eigenvalues ​​of each covariance matrix are solved, and the eigenvector corresponding to the minimum eigenvalue is selected from the eigenvalues ​​of each covariance matrix as the normal vector corresponding to each neighborhood. The recognition system obtains the first number of inner points, the second number of inner points and the third number of inner points according to the normal vectors after dimensionality reduction, and judges whether the first number of inner points, the second number of inner points and the third number of inner points meet the corresponding threshold conditions respectively. If the first number of inner points, the second number of inner points and the third number of inner points all meet the corresponding threshold conditions respectively, the current scene is determined to be a long corridor environment.

The invention relates to the technical field of automatic driving, in particular to an automatic driving control method and device based on laser ranging, a vehicle and a medium, and the method comprises the steps: obtaining a current automatic driving working condition of the vehicle, matching a target ranging mode of a mode-locked pulse laser and at least one ranging target of a current environment, and generating a measurement action of at least one distance measurement target according to the target distance measurement mode, executing the measurement action by using the double-optical comb pulse laser output by the mode-locked pulse laser until the current distance data of the at least one distance measurement target is obtained, exiting the target distance measurement mode, and controlling the vehicle to run in the current automatic driving working condition according to the current distance data. According to the embodiment of the invention, the distance measurement task of the current working condition of the vehicle can be executed by applying the double-optical-comb pulse laser output by the mode-locked pulse laser, so that the target distance of the vehicle is accurately measured, the accuracy and reliability of laser data are improved, and the driving safety of a user is further guaranteed.

This invention discloses a cliff detection method and terminal. Single-point lasers are installed downwards directly in front of and behind the left and right drive wheels of a vehicle, respectively. A single-point laser is also installed at an angle downwards directly in front of the vehicle. The cliff situation near the drive wheels is measured using these single-point lasers. The angled single-point laser at the front of the vehicle avoids the laser detection position being too close to the front wheels when the vehicle is moving, preventing the vehicle from failing to stop in time due to inertia. The method selects the appropriate single-point laser group for detection based on the vehicle's motion state. Therefore, irrelevant single-point laser data does not participate in the logic under different motion states, reducing the possibility of false signal triggering. If the distance detected by one of the single-point lasers in the single-point laser group exceeds a preset distance, the vehicle braking program is triggered, and the vehicle moves in the direction of the single-point laser whose detected distance is within the preset distance. This accurately assesses the surrounding ground environment and improves the accuracy of cliff detection.

The invention relates to a guide rail support spacing adjusting device and method for an elevator, and belongs to the technical field of elevator installation and maintenance, and the guide rail support spacing adjusting device comprises a measuring trolley climbing along a guide rail, a short-range laser sensor array driven by an actuator, a foresight distance sensor and a tilt angle sensor; laser, stroke and attitude data are synchronously collected during climbing; detecting the guide rail bracket by a foresight sensor; when the guide rail bracket is detected, recording an initial position and withdrawing the laser array; after crossing the bracket, recording a termination position and recovering the array; correcting the laser data according to the attitude data, and interpolating and reconstructing a guide rail contour of a support shielding area to generate an adjusting scheme; coevolution of measurement and avoidance is achieved, the existence of the foresight distance sensor provides a pre-judgment capability for a control algorithm, the avoidance action can be carried out actively in advance, and the method is a physical basis for achieving intellectualization of the algorithm.

The application relates to the technical field of target detection, and discloses a detection method and device for small targets on a road during automatic driving, which comprises the following steps: obtaining original laser data of an intelligent automobile during driving by using a laser module, and shooting original shooting pictures during driving by using a camera module; performing first target detection on the original shooting pictures to obtain a first target detection frame; obtaining a target to-be-corrected point set by taking each pixel point in the first target detection frame as a target to-be-corrected point; converting original laser data belonging to a laser coordinate system into a focal length coordinate system by using a camera coordinate system to obtain focal length laser data; cleaning the target to-be-corrected point set by using the focal length laser data to obtain a corrected detection frame; and replacing the first target detection frame with the corrected detection frame to obtain a target detection result during automatic driving of the intelligent automobile. The main purpose of the application is to solve the problem that a target detection frame generated by detecting small targets during automatic driving is inaccurate.

This invention discloses a dynamic environment SLAM localization method based on multi-source sensor data fusion, comprising the following steps: acquiring visual, inertial, and laser data and calibrating the buffer; performing HRNet segmentation and ORB-SLAM tracking on the visual data to obtain the initial pose; fusing the inertial and laser results to obtain the current pose; determining and recalculating the divergence of time window data to update the current pose; inserting keyframes and generating a semantic skeleton map; encoding and generating a semantic topology string and associating it with the keyframe index; retrieving loop closure candidate frames and performing geometric verification followed by map optimization to obtain the localization result and map result. This invention achieves highly robust localization and mapping in dynamic environments and improves the accuracy of loop closure recognition and map consistency.