9results about How to "Accurate pose" patented technology

The application discloses a real-time fine mapping method suitable for an underwater robot, which comprises the following steps: firstly, a 3D Gaussian map is constructed by using a first frame; then, key frame determination is performed on the following image frames; and finally, an incremental map is constructed by using the key frames; the process effectively overcomes the optical interference specific to the underwater environment, filters out the geometric noise caused by suspended particulate matters and the like, and significantly improves the geometric consistency and color authenticity of the map; all frames are tracked in real time, accurate poses are obtained and output, the process effectively overcomes the interference of underwater optical degradation on pose estimation, and ensures the robustness and accuracy of the tracking process.

The present disclosure relates to a method and device for determining camera pose, and an electronic device, and particularly relates to the technical field of visual positioning and tracking. The method comprises: acquiring a target image captured by a camera; determining a plurality of first feature points in voxels; when there is no other feature point in the voxels between a first target feature point and a center point of the camera, performing feature point matching between the first target feature point and a pixel point in the target image, the first target feature point being any one of the plurality of first feature points; and if the first target feature point matches the pixel point in the target image, calculating an actual pose of the camera based on the first target feature point.

The present disclosure provides a pose determination method, a pose determination device, a computer readable storage medium and an electronic device, and relates to the technical field of computer vision. The pose determination method comprises the following steps: determining first two-dimensional feature points on a current frame color image collected by a first camera and matched with a previous frame color image collected by the first camera, determining second two-dimensional feature points on a current frame color image collected by a second camera and matched with a previous frame color image collected by the second camera, converting the second two-dimensional feature points to a first camera coordinate system to obtain third two-dimensional feature points, and determining a pose of the first camera when collecting the current frame color image according to the first two-dimensional feature points, the third two-dimensional feature points, three-dimensional feature points of the previous frame color image collected by the first camera in a world coordinate system, and three-dimensional feature points of the previous frame color image collected by the second camera in the world coordinate system. The present disclosure can improve the positioning accuracy.

The application relates to a visual positioning method, a device and an intelligent vehicle. The method comprises the following steps: acquiring an image including complete ground road signs collected by a camera, segmenting the image including the complete ground road signs to obtain a binary image, extracting corner points of real ground road signs in the binary image, acquiring historical poses of the camera and a ground road sign map, obtaining estimated corner points of the ground road signs according to the historical poses and the ground road sign map, matching the estimated corner points of the ground road signs with the corner points of the real ground road signs, and obtaining a pose of the camera at present. The application aims to solve the positioning drift phenomenon of a visual positioning system of an intelligent vehicle during driving, match the estimated corner points of the ground road signs with the corner points of the real ground road signs, continuously correct the pose deviation of a real camera and an estimated camera, make the obtained pose of the camera at a current time more accurate, and improve the positioning accuracy of the intelligent vehicle.

The embodiment of the invention is suitable for the technical field of smart home, and provides a control method and device of a sweeping robot, the sweeping robot and a storage medium, and the method comprises the steps: collecting environment point cloud data in a current direction through a distance measuring sensor of the sweeping robot; environment point cloud data intersection information is determined according to the environment point cloud data and a currently established map block of the sweeping robot; and according to the environment point cloud data intersection information, controlling the sweeping robot to execute a preset action so as to collect environment point cloud data in multiple directions. Through the method, the sweeping robot can be rotated in time, so that the scanning view angle of the distance measuring sensor of the sweeping robot is adjusted, and the quality of the built map blocks of the sweeping robot is optimized.

This invention relates to the field of optical alignment technology, specifically providing a global field-of-view calibration method and device based on a digital model of an alignment system. The method includes: establishing a measured coordinate system and acquiring measured point cloud data; matching the measured point cloud data with ideal point cloud data to determine the direction of the optical axis in the measured coordinate system; constructing a detection optical path and acquiring the mirror model of the standard mirror in the detection optical path and three reference planes on the standard mirror fixture; calculating the normal vector of the standard mirror and the spatial pose relationship between the reference planes and the standard mirror; establishing a field-of-view coordinate system and transforming the spatial pose relationship between the reference planes and the standard mirror to the field-of-view coordinate system; and calculating the field-of-view angle of the normal vector of the standard mirror. This invention reliably transmits the optical surface pose, which cannot be directly measured, to a field-of-view coordinate system established based on the secondary mirror of the camera under test, making the calculation of the optical axis direction and field-of-view angle intuitive and accurate, effectively ensuring the alignment accuracy of asymmetric, off-axis optical systems such as off-axis three-lens reflex cameras.

The present application relates to a kind of based on multi-sensor fusion's heading machine tracking positioning method, belong to heading machine positioning technical field.The method includes: fixedly installing total station in the roadway top behind the heading direction of heading machine, suspending tracking positioning robot in the guide rail above roadway, carrying binocular camera, strapdown inertial navigation system on tracking positioning robot, installing infrared target plane on the body of heading machine;Control tracking positioning robot moves to the best observation distance interval of binocular camera, and obtains the target image of infrared target plane collected by binocular camera;According to target image, calculate camera-carrier rotation matrix and camera-carrier translation matrix;Obtain the attitude of tracking positioning robot relative to navigation coordinate system;According to camera-carrier rotation matrix, camera-carrier translation matrix, the attitude of tracking positioning robot relative to navigation coordinate system and coordinate system conversion relationship determine the pose of heading machine.The present application can continuously track and locate heading machine.

Embodiments of the present application disclose a target object pose determination method and device and electronic equipment. The method performs coarse and fine granularity matching on the geometric features of the scene point cloud and the geometric features of the model point cloud, respectively establishes a first matching result and a second matching result between the scene point cloud and the model point cloud, and then calculates a target pose of the target object based on the first matching result and the second matching result. Through the present application, the technical problem that the poses of all objects in an application scene cannot be accurately estimated and it is difficult to meet the requirements of an application scene with high pose accuracy is solved, and the technical effect of more accurately identifying the pose of an object and improving the adaptability of a complex application scene is achieved.

The application discloses a visual odometer method fusing mid-term semantic information, comprising the following steps: S1. a tracking thread receives a binocular image stream, and delivers a monocular image to a semantic segmentation process; the semantic segmentation process performs semantic segmentation on the monocular image, and delivers a semantic segmentation graph to the tracking thread; S2. the tracking thread constructs a current frame according to the semantic segmentation graph and generates an initial map; local map points are matched with feature points of the current frame; the pose of the current frame is optimized; a key frame is determined; key frame information and the semantic segmentation graph are transmitted to a local mapping thread; S3. the local mapping thread updates map points according to the feature points of the key frame, constructs a mid-term sliding window and updates semantic map points of the mid-term sliding window, establishes a mid-term constraint according to the semantic map points and optimizes the pose of the key frame; S4. the local mapping thread reversely projects the map points to the semantic segmentation graph by using the optimized pose, updates semantic information of the semantic map points, and manages the semantic map points according to the quality of the semantic map points.