42 results about "Map projection" patented technology

In various examples, detecting occluded objects within images or other sensor data representations for autonomous or semi-autonomous systems and applications is described herein. Systems and methods described herein may determine when objects are occluded at portions of images using various techniques. For example, an image may be processed in order to determine classifications associated with objects depicted by the image and, the classifications, along with labels that are projected on the image using a map, may then be used to determine whether one or more of the objects are occluded in the image. For another example, a map may be used to determine first distances to points within an environment and a point cloud may be used to determine second distances to the points within the environment. The distances may then be used to determine whether one or more objects are occluded within the image.

In various examples, detecting occluded objects within images or other sensor data representations for autonomous or semi-autonomous systems and applications is described herein. Systems and methods described herein may determine when objects are occluded at portions of images using various techniques. For example, an image may be processed in order to determine classifications associated with objects depicted by the image and, the classifications, along with labels that are projected on the image using a map, may then be used to determine whether one or more of the objects are occluded in the image. For another example, a map may be used to determine first distances to points within an environment and a point cloud may be used to determine second distances to the points within the environment. The distances may then be used to determine whether one or more objects are occluded within the image.

The invention discloses a foggy day driving augmented reality auxiliary method and system based on multi-mode perception, and relates to the technical field of intelligent driving. The method comprises the following steps: constructing a three-dimensional environment map to carry out real-time driving risk prediction, and carrying out preliminary adjustment on data source weight distribution in the three-dimensional environment map according to a set rule; the PPO reinforcement learning algorithm is adopted to readjust data source weight distribution, a new three-dimensional environment map is reconstructed based on the readjusted data source weight, and the risk probability, predicted in real time, of sudden deceleration of the front vehicle and the reconstructed three-dimensional environment map are projected to a windshield through a layered projection strategy to prompt a driver. According to the method, accurate and comprehensive multi-mode sensing data can be obtained in a foggy day environment, the accuracy of decision making and risk prediction is ensured, quick response can be made according to environment changes, the safety of a driver under complex weather conditions is improved, and the method can adapt to more changing environments in practical application.

The invention discloses a tight coupling laser inertial vision fusion method based on radiation reconstruction, and relates to the technical field of vision fusion. Comprising the following steps: removing laser radar point cloud motion distortion based on IMU measurement, dividing the laser radar point cloud into multi-scale voxel grids, and obtaining a plane normal vector of the laser radar point cloud contained in each voxel grid to calculate a point-surface residual error; real radiation information is obtained through photometric calibration, meanwhile, a global map is projected to a current frame to obtain a tracking point, the obtained tracking point is tracked through an LK optical flow method, and a frame-to-frame re-projection error and a radiation error are obtained; a multi-constraint factor graph containing IMU pre-integration, laser radar constraint and visual constraint is constructed based on the factor graph, and robustness enhancement of the system in a degradation environment is realized in combination with a key frame sliding window optimization mechanism. The method provided by the invention has high positioning precision and robustness in a highly challenged degradation scene.

The invention discloses a tight coupling laser inertial vision fusion method based on a merged probability voxel map, and relates to the technical field of vision fusion. Comprising the steps of performing display parametric modeling based on radar measurement noise, and constructing a probability voxel plane model; combining the planes which may have a coplanar relationship, and adding the laser radar point clouds in the determined effective voxel planes into a global map; projecting the global map into an image frame to obtain a tracking point, tracking by using an LK optical flow method, and meanwhile, further eliminating an abnormal point by using a random sampling consensus algorithm based on a dynamic Bayesian network; and based on image information acquired by the camera, optimizing and maintaining the global map by minimizing frame-to-frame pixel errors and frame-to-map color errors and updating the global map. According to the method, the performance of the laser inertial visual odometer is improved by combining a framework of efficient coupling radar, IMU and camera information of the probability voxel map.

The application discloses a kind of unmanned aerial vehicle image and multi-modal map area cutting method and system based on attitude information, first obtain the unmanned aerial vehicle image containing EXIF metadata, extract GPS information and attitude parameter, calculate image ground real projection center point longitude and latitude in combination with sensor parameter graph;Again, according to pixel size, sensor physical size etc., calculate ground resolution and geographical coverage, correct the projection angle to get four real longitude and latitude, then four corner coordinates are converted to map projection coordinate system, generate cutting window in combination with base map geographic reference information, high-precision cutting is carried out in infrared, visible light map base map respectively, corresponding result is carried out after spatial alignment and consistency verification, and relevant data is saved in JSON format structure.The method realizes accurate area matching and cutting, improves the spatial accuracy and processing efficiency of multi-source remote sensing data fusion and target identification, and is suitable for remote sensing surveying and mapping, resource investigation and emergency monitoring and other scenes.

The invention belongs to the technical field of astronomical image data processing, particularly relates to an astronomical image data set labeling method, and aims to solve the problem that existing single threshold segmentation-based labeling and star map projection-based labeling are difficult to meet the requirement of a depth model for a high-quality data set. According to the method, on the basis of threshold segmentation labeling and star map projection labeling, a first labeling result and a difference set are adopted as a labeling result A for an observation image large-probability detectable target, a second labeling result is adopted as a labeling result B for an observation image small-probability detectable target, and the labeling result A and the labeling result B are integrated to obtain a labeling result A and a labeling result B; and obtaining a final labeling result. According to the method, the integrity of target labeling can be improved, the probability of missing labeling and wrong labeling can be reduced, the actual fitting degree of the labeling result and the image can be met, unification of integrity and accuracy is achieved, and the depth model performance can be improved through the manufactured data set.

Disclosed are a wide viewing angle stereo camera apparatus and a depth image processing method using the same. A stereo camera apparatus includes a receiver configured to receive a first image and a second image of a subject captured through a first lens and a second lens that are provided in a vertical direction; a converter configured to convert the received first image and second image using a map projection scheme; and a processing configured to extract a depth of the subject by performing stereo matching on the first image and the second image converted using the map projection scheme, in a height direction.

The invention discloses an unmanned aerial vehicle image and multi-modal map area cutting method and system based on attitude information, and the method comprises the steps: firstly obtaining an unmanned aerial vehicle image containing EXIF metadata, extracting GPS information and attitude parameters, and calculating the latitude and longitude of a real projection center point of the image ground in combination with sensor parameters; calculating ground resolution and geographical coverage according to pixel size, sensor physical size and the like, correcting projection angles to obtain real longitudes and latitudes of four corners, transferring coordinates of the four corners to a map projection coordinate system, generating a cutting window by combining geographical reference information of the base map, and performing high-precision cutting in the base map of the infrared and visible light map respectively to obtain a three-dimensional image; after a corresponding result is obtained, space alignment and consistency verification are carried out, and related data are stored in a JSON format in a structured mode. The method realizes accurate region matching and cutting, improves spatial precision and processing efficiency of multi-source remote sensing data fusion and target identification, and is suitable for various scenes such as remote sensing surveying and mapping, resource investigation, emergency monitoring and the like.

This application provides a map registration method and apparatus. In this application, after acquiring a first map of the target dimension and a two-dimensional map for registering the first map, a reference coordinate system for the first map is determined, and the two-dimensional map is projected onto this reference coordinate system to obtain a second map. This ensures that the first and second maps are in the same reference coordinate system, and that there is a correspondence between the coordinates of points in the first and second maps. Based on this, the first map can be registered using the coordinate correspondence between points in the first and second maps, as well as the geographical location information of the registration reference points in the second map. The solution of this application can improve the accuracy of map registration.

Various examples describe the detection of occluded objects in images or other sensor data representations for autonomous or semi-autonomous systems and applications. The systems and methods described herein may determine when objects in portions of images are occluded using various techniques. For example, an image may be processed to determine classifications associated with the objects represented in the image, and the classifications may then be used, along with labels projected onto the image using a map, to determine whether one or more of the objects in the image are occluded. As another example, a map may be used to determine initial distances to points within an environment, and a point cloud may be used to determine secondary distances to the points within the environment.The distances can then be used to determine whether one or more objects within the image are occluded.

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 invention relates to a multi-robot relative positioning method for a field environment, and the method comprises the steps: constructing a local point cloud map, removing ground point cloud, obtaining a non-ground local point cloud map, projecting the non-ground local point cloud map to a bird's-eye view, and obtaining a multi-frame bird's-eye view space height distribution diagram as a historical frame; screening a candidate loopback frame; screening out an effective loopback frame from the plurality of candidate loopback frames, and obtaining a relative pose of the robot between the current frame and the effective loopback frame as a loopback constraint; removing abnormal values in the loopback constraint set to obtain a reliable loopback constraint set; loopback factors are generated based on the reliable loopback constraint set, and a joint factor graph is constructed; and carrying out adaptive weight adjustment on the loopback factors in the joint factor graph, removing the loopback factors of which the weights are lower than a threshold value, updating the joint factor graph, and updating pose estimation results and global consistent trajectory estimation results of all robots at the same time.

The present invention discloses an on-orbit geometric positioning method and system for a high-orbit large area array camera based on the earth's contour, which relates to the field of space remote sensing technology. An on-orbit geometric model of the high-orbit large area array camera is established; the earth's contour within the observation field of the model camera is extracted to obtain the coordinates of the earth's contour; based on the coordinates of the earth's contour, the center point of the earth's contour is obtained; the camera attitude angle is calculated according to the position of the center point of the earth's contour and the center of the earth; the polar coordinates of the remote sensing image in the LOS projection plane are calculated using the azimuth angle and elevation angle of the camera attitude angle, and the projection of the remote sensing image during reference map projection matching is obtained using the rotation angle of the camera attitude angle, and the transformation of the image point in the remote sensing image to the geocentric fixed system coordinates of the corresponding earth surface point is carried out to achieve on-orbit geometric positioning. The on-orbit geometric positioning method and system for a high-orbit large area array camera based on the earth's contour provided by the present invention solve the technical problems that factors such as clouds, the number and distribution of control points have a great influence on the positioning accuracy.

A map updating method, comprising: acquiring environmental point cloud data collected at a target pose in a robot operation area, and acquiring an initial local three-dimensional map corresponding to a robot sensing space corresponding to the target pose (S202); in the initial local three-dimensional map, removing sensing information corresponding to the environmental point cloud data in the robot sensing space to obtain an intermediate local three-dimensional map corresponding to the robot sensing space (S204); mapping the environmental point cloud data onto the intermediate local three-dimensional map to obtain a target local three-dimensional map corresponding to the robot sensing space (S206); and mapping the target local three-dimensional map onto an initial two-dimensional map corresponding to the robot operation area to obtain a target two-dimensional map (S208).

Systems and methods for autonomous vehicle path planning are described herein. An example vehicle includes an image sensor to obtain an image of a scene of an area surrounding the vehicle. The vehicle also includes navigation system circuitry to: analyze the image and generate a semantically segmented image that identifies one or more types of features in the image; project the semantically segmented image to a two-dimensional (2D) map projection; convert the 2D map projection into a cost map; and determine a path for the vehicle based on the cost map.

The application discloses a vehicle logo positioning method based on visual saliency and binary map projection segmentation, and comprises the following steps: positioning a license plate, obtaining a vehicle logo rough positioning map according to the positions of four corner points of the license plate; performing vehicle logo accurate positioning on the vehicle logo rough positioning map by using visual saliency and binary map projection segmentation in parallel, and obtaining a candidate vehicle logo vector and a vehicle logo positioning rectangular vector respectively; generating a vehicle logo candidate rectangular vector based on the candidate vehicle logo vector and the vehicle logo positioning rectangular vector; and filtering non-vehicle logos in the vehicle logo candidate rectangular vector by using a vehicle logo-non-vehicle logo sorting model based on deep learning, and generating a vehicle logo rectangular vector. The application performs vehicle logo rough positioning based on visual saliency, then positions the vehicle logo according to an empirical value, simultaneously detects a gray peak value area of a vehicle logo visual saliency map by using a binary map projection segmentation method, supplements a map in which visual saliency positioning fails, overcomes the shortcoming that a single method is difficult to cope with variable environments, and finally filters non-vehicle logos by using a vehicle logo-non-vehicle logo sorting model, so that the success rate of vehicle logo positioning is superior to that of an existing method.

The present disclosure relates to a hemi-cube map projection format in an imaging environment. According to one embodiment, mechanisms for facilitating a hemi-cube map projection format imaging environment are described. The method of embodiments as described herein includes capturing, by a camera coupled to one or more processors, an image having image content, wherein the image content represented by the image is omnidirectional such that the image content, while representing less than a sphere, is mapped on a sphere; mapping the image to a cube representation based on six faces of the cube, wherein one or more of the six faces are classified as inactive regions such that the one or more faces remain unoccupied or partially occupied; and arranging the six faces in a compact representation based on the cube representation by avoiding inclusion of the inactive regions.

The application discloses a two-three-dimensional map adaptive projection matching method based on a hidden Markov chain, constructs a hidden Markov projection prediction model based on user preferences and projection characteristics, takes a map projection category as a hidden state, takes a user browsing position or a visual angle as an observation state, fuses user preferences and projection standard parameters to form a probability of each hidden state occurrence at each moment, updates an observation probability at each moment, and forms a current optimal projection matching combination through continuous iteration. The application integrates quantitative projection characteristics, combines user preferences, and influences a two-three-dimensional GIS visualization mode in a brand-new concept, selects a map projection suitable for different users and different scenes from an intelligent perspective, effectively overcomes browsing obstacles caused by a lack of projection field knowledge of part of users, and enhances a user's visualization experience.

The present application discloses a method, apparatus, and electronic device for determining vehicle position and attitude data, which relates to the field of computer technology, and particularly to artificial intelligence fields such as autonomous driving, vehicle networking, intelligent cockpits, and computer vision. The specific implementation solution is as follows: Obtain the first position and attitude data of the vehicle at the first moment, the first predicted value of the first relative state of the vehicle at the second moment relative to the first moment, and the first error value of the error variable corresponding to the first relative state; wherein, the first moment is earlier than the second moment; Determine the first estimated value of the first relative state according to the first error value and the first predicted value; Determine the second position and attitude data of the vehicle at the second moment according to the first position and attitude data and the first estimated value. This method determines the position and attitude data of the vehicle at the second moment based on the error value of the error variable of the relative state between different moments, thereby being unaffected by map projection errors and improving the accuracy of the position and attitude data.

This invention discloses an automatic obstacle avoidance method for exterior wall spraying robots based on lidar and BIM, belonging to the field of exterior wall spraying robot technology. It solves the problem that existing exterior wall spraying robots cannot automatically avoid obstacles such as windows or balconies. This invention obtains a point cloud map of the exterior wall through a BIM model and projects the point cloud map into a grid map for planning operations. Based on a two-step point cloud registration method using feature extraction, the lidar point cloud is registered with the point cloud map. This not only realizes the localization of the exterior wall spraying robot, but also obtains information on obstacles such as windows and balconies. This not only improves the accuracy and efficiency of spraying operations, but also reduces human intervention and lowers operational risks, especially in high-altitude operations, greatly improving safety.

The invention provides a mining area point cloud map projection method and device and electronic equipment, and relates to the technical field of mine unmanned driving. The method comprises the following steps: acquiring local point cloud window data to be registered and global reference pose data; the local point cloud window data is window data of mining area local point cloud generated based on simultaneous localization and mapping (SLAM); determining a target time offset between the local point cloud window data and the global reference pose data; performing registration processing on the local point cloud window data and a global reference map according to the target time offset and the global reference pose data to generate a corresponding accurate transformation matrix; and projecting the local point cloud window data to the global coordinate system according to the accurate transformation matrix, and generating corresponding target point cloud map data. According to the method, global point cloud map data for global navigation and scheduling can be generated. And meanwhile, the map projection efficiency is also improved.

The invention relates to a three-dimensional teaching method, device and system for map projection, and relates to the technical field of map projection teaching. According to the method, the device and the system, the three-dimensional model and the three-dimensional animation are displayed by setting the display equipment, so that teaching is more vivid; the eye tracking device and the handheld interaction device are arranged, interaction rules are defined, interactive teaching is achieved, the participation degree of students in the learning process is improved, and the immersion feeling and experience feeling of the students in the learning process are improved.

An object of the present invention is to provide an information processing method, a program, and an information processing device that can improve the quality of a composite image. [Solution] An information processing method for generating a composite image by synthesizing hand-drawn images is provided, the method comprising an image acquisition step, an image conversion step, and an image synthesis step, wherein the image acquisition step acquires a first image and a second image, the first image being an image represented in cube map projection and the second image being an image represented in equirectangular projection, the image conversion step converts the first image from an arrangement in the cube map projection to an arrangement in the equirectangular projection to generate a transformed first image, and the image synthesis step synthesizes at least the transformed first image and the second image to generate the composite image.

This invention relates to the field of lawn mowing robot technology, and more particularly to a method for improving the single-point GNSS positioning accuracy of lawn mowing robots by integrating boundary information. The method includes the following steps: S1, calculating the coordinates of the lawn boundary; S2, constructing a state transition probability using the difference between the displacement vector of the map projection point and the displacement vector of the trajectory point; S3, if the number of state transitions that do not meet the constraints in the most recent n state transitions is greater than m, then it is determined that an obstacle boundary has been detected; S4, if the number of obstacle boundary maps equals the number of obstacle boundaries, modifying the state transition probability to the original equation; S5, constructing the coordinates of the obstacle boundary by using the difference between the position and heading of the starting trajectory point around the obstacle boundary and the current trajectory point. This invention solves the problem of how robots can independently acquire map information, and the long-term drift of single-point GNSS positioning, which leads to repeated mowing or missed mowing of lawns during lawn mowing.

A computer includes a processor and a memory, and the memory stores instructions executable by the processor to generate an observed overhead feature map from an overhead image of a geographic area; generate a ground-view feature map from a ground-view image captured by a camera within the geographic area, the camera oriented at least partially horizontally while capturing the ground-view image; project the ground-view feature map to an overhead view based on a geometric relationship, resulting in a projected overhead feature map; determine a rotation of the camera relative to the overhead image based on the projected overhead feature map and the observed overhead feature map; and determine a translation of the camera relative to the overhead image based on the projected overhead feature map and the observed overhead feature map.

The invention belongs to the technical field related to 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 a covariance matrix of neighborhood point clouds of each feature point; s2, solving the corresponding optimal attitude when the objective function is minimum; s3, converting the obtained point cloud data into a world coordinate system by using the optimal attitude, fusing all frames of 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 a semantic segmentation neural network; and obtaining a semantic label corresponding to each point cloud in the three-dimensional point cloud map so as to realize semantic mapping. According to the invention, the problems of weak dynamic adaptability, insufficient model generalization ability and long-term operation precision attenuation of a traditional three-dimensional semantic mapping method in a dynamic complex scene are solved.

Methods, systems, devices, and non-transitory computer readable media for determining landmark visibility and generating annotations are provided. The disclosed technology can include receiving map data comprising three-dimensional models of structures in a physical environment. Portions of the three-dimensional models of structures that are visible from map projection cells associated with the physical environment can be determined. Visibility data associated with portions of the three-dimensional models that are visible from each of the map projection cells and comprising visibility cells corresponding to the map projection cells can be generated. View data associated with a visual representation of the physical environment from a location can be received. Based on the visibility data and view data, the visibility cell associated with the location in the physical environment can be determined. Based on point of interest data, points of interest associated with the visibility cells can be determined and annotations can be generated.

In various examples, detecting occluded objects within images or other sensor data representations for autonomous or semi-autonomous systems and applications is described herein. Systems and methods described herein may determine when objects are occluded at portions of images using various techniques. For example, an image may be processed in order to determine classifications associated with objects depicted by the image and, the classifications, along with labels that are projected on the image using a map, may then be used to determine whether one or more of the objects are occluded in the image. For another example, a map may be used to determine first distances to points within an environment and a point cloud may be used to determine second distances to the points within the environment. The distances may then be used to determine whether one or more objects are occluded within the image.

The embodiment of the invention provides a motor vehicle mapping and positioning method and device and a computer readable storage medium. The method comprises the following steps: obtaining a look-around spliced two-dimensional diagram based on video images of multiple paths of vehicle-mounted cameras; performing semantic segmentation on the road marks in the look-around spliced two-dimensional graph to obtain line segment template images and parking space template images; performing image processing on the line segment template image and the parking space template image to obtain an initial two-dimensional semantic image; performing pixel clustering and parameterization processing on the initial two-dimensional semantic image to obtain a target two-dimensional semantic image; projecting the target two-dimensional semantic image to a vehicle coordinate system based on the speedometer pose to obtain an initial local map, and clustering and fusing continuous multiple frames of initial local maps to obtain a precise local map; projecting the accurate local map to a longitude and latitude coordinate system to obtain a target local map, and further obtaining a lightweight road semantic map; and matching and positioning the motor vehicle. According to the embodiment of the invention, mapping and positioning can be realized in a scene with less textures.