23 results about "Collinearity equation" patented technology

The invention provides a light and small unmanned aerial vehicle push-scan image distortion correction method, and relates to the technical field of unmanned aerial vehicle image correction, and the method comprises the following steps: erecting an RTK base station on a known point, and building a mobile workstation; the push-scan image of the unmanned aerial vehicle is obtained through a hyperspectral imager carried by an unmanned aerial vehicle holder; the unmanned aerial vehicle sends record data and push-scan image data of the carried POS to the mobile workstation; performing preliminary geometric correction through a collinear equation method; performing ortho-rectification by combining DEM topographic data; and sub-pixel-level correction is carried out in combination with an SIFT algorithm and an FFT method. The method has high efficiency and robustness for medium noise and lens optical blurring, and the problem of image distortion caused by rotation effects such as inclination, pitching and rolling in the flight process can be solved.

The invention discloses a cultivated land non-agrochemical monitoring method based on unmanned aerial vehicle video spatialization. The method comprises the following steps: acquiring a farmland area video stream, positioning and attitude determination system data and digital elevation model data acquired by an unmanned aerial vehicle; extracting a key frame from the video stream based on a double-threshold detection algorithm of the optical stream change rate and the image information entropy; performing spatialization processing on the key frame through a collinear equation optimization solution combined with a digital elevation model, and converting image coordinates into geographic coordinates; fusing the key frame image data with the corresponding geographical coordinate information, and constructing input data of a deep learning model of geographical perception; inputting the fused data into a geographic perception deep learning model to perform ground object interpretation, and extracting spatial features by the model through a built-in geographic coordinate embedding layer to identify a cultivated land non-agrochemical target; post-processing is carried out on the interpretation result, the post-processing comprises morphological closed operation processing and grid-to-vector conversion processing, quality control is carried out by calculating a re-projection error index and an intersection-to-parallel ratio index, and a cultivated land non-agrochemical monitoring report is generated.

The invention discloses a three-dimensional human body posture recognition method based on photogrammetry and deep learning, and relates to the technical field of posture recognition, and the method comprises the following steps: collecting original video data, carrying out the posture estimation through employing a three-dimensional posture estimation method MediaPipe based on deep learning, and obtaining the original three-dimensional key point data of an image; carrying out preliminary constraint on the key point data by using a key point specification algorithm, and correcting an identification error; performing data enhancement by using a multi-view simulation algorithm based on a collinear equation to obtain richer view information and expand a data set; the method comprises the following steps: constructing a posture recognition model KeyposeNet; and carrying out training and testing on the KeyposeNet, and selecting an optimal model to carry out posture recognition. According to the three-dimensional human body posture recognition method based on photogrammetry and deep learning, more efficient and accurate human body posture classification recognition is achieved, the cost of manual classification is reduced, an auxiliary means is provided for human body posture recognition and judgment, and therefore the three-dimensional human body posture recognition method based on photogrammetry and deep learning is applied to various occasions.

The application discloses a kind of antenna profile rapid measurement system real-time calibration method, comprising: each laser tracker respectively to each control point is aimed at ranging, the distance value of each laser tracker center and each control point is obtained, create space sphere, obtain the accurate coordinate of each control point;Antenna profile rapid measurement system is photographed to the control point and the measured mark point of measured antenna imaging, the image point coordinate of control point and the image point coordinate of the measured mark point of measured antenna are obtained;Based on control point collinear condition equation and the preset measurement camera internal parameter, the position and posture of control point relative to control field coordinate system are obtained according to the image point coordinate of control point and the image point coordinate of the measured mark point of measured antenna;According to profile measurement point collinear equation and the position and posture of control point relative to control field coordinate system, the position and posture of the measured mark point of measured antenna are obtained.The application improves the measurement accuracy and reliability of antenna profile rapid measurement system.

The application provides a forest and grassland fire early identification and positioning method based on infrared thermal imaging and visible light fusion, and belongs to the technical field of forest and grassland resource protection and ecological safety monitoring. The method comprises the following steps: synchronously collecting infrared and visible light images and recording the space-time information of a monitoring platform; performing pixel-level registration on the two images; extracting a thermal anomaly suspected area from the infrared image based on a dynamic background temperature field, and extracting a smoke suspected area from the visible light image based on color, texture and motion features; performing spatial correlation analysis on the suspected areas extracted from the two channels, and combining the area growth rate and temperature rise rate of continuous multiple frames to construct a fire evolution model to confirm a fire point; calculating the three-dimensional geographic coordinates of the fire point based on the photogrammetry collinear equation and a digital elevation model; and generating and outputting alarm information containing the three-dimensional coordinates and fire intensity. The application realizes early and accurate identification and positioning of forest and grassland fires, and significantly reduces the false alarm rate.

The present invention discloses a tidal flat terrain monitoring method based on few-shot learning. This method utilizes the non-orthogonal tidal flat flooding image data collected by the coastal monitoring system; with the support of a small number of labeled samples, through introducing the large-scale basic model SAM for reinforcement learning, it realizes the rapid and accurate extraction of the tidal flat waterline; uses the collinearity equation to convert the non-orthogonal waterline into an orthogonal waterline; combined with the real-time tidal level information provided by the water level monitoring station, this technology can construct a high-precision tidal flat terrain on the time scale of the tidal cycle. The advantages of the present invention are: it provides an effective solution strategy for the tidal flat terrain monitoring field with scarce labeled data, and further supports the establishment of a fully automated and dynamic monitoring system for the tidal flat terrain.

The present disclosure provides a real-time processing system and method for active and passive three-dimensional imaging, including: a PS end, connected to a peripheral end through an external interface controller, using a serial processor to implement low-parallel operations, the PS end including: a system initialization module; a data preprocessing module; a laser ranging data error correction module; an optical image data processing module; a PL end, connected to the PS end through an AXI bus, using programmable logic to implement highly parallel operations; the PL end including: two DMA controllers, a first DMA controller and a second DMA controller; a geodetic coordinate solution module, connected to the first DMA controller; a Gaussian projection module, one end of which is connected to the geodetic coordinate solution module; and a collinear equation solution module, connected to the second DMA controller. The active and passive three-dimensional imaging processing methods and processes are reasonably divided.

The invention discloses a railway facility inspection method and system fusing GNSS positioning and image recognition, and relates to the technical field of rail traffic monitoring, and the method comprises the steps: employing a GNSS receiver to obtain a second pulse signal, triggering an inertial measurement unit, an encoder and a line-scan digital camera, and obtaining the real-time attitude data and image data of a track along the line; carrying out tight coupling calculation on the real-time attitude data, generating a mobile station track with centimeter-level precision, obtaining exterior orientation elements of image data shot by a linear array camera, mapping pixel coordinates of the image data to a geodetic coordinate system by adopting a collinear equation, calibrating latitude and longitude and mileage information for each abnormal target pixel, and obtaining fused image data; inputting the fused image data into the trained deep learning network for processing, and obtaining a recognition result; and comparing the identification result with preset geometric data, and if an abnormal condition is satisfied, obtaining a final identification result. The method can improve the confidence of the image recognition result.

The invention discloses a distributed vision measurement method and system applied to multiple large fishponds. The method comprises the following steps: arranging a measuring station near each fishpond to be measured, and obtaining a left photo and a right photo; constructing an optical distortion model of the camera lens, respectively calculating distortion values of the plurality of control points and the plurality of pointer points based on the optical distortion model, carrying out descending sorting on the distortion values, and screening a preset number of control points and pointer points with minimum distortion values; constructing a collinear equation, and resolving the collinear equation through the screened control points to obtain exterior orientation elements; and forward intersection calculation is carried out based on the interior orientation element and the exterior orientation element of the camera and the image point coordinate calculation of the to-be-detected key point of the to-be-detected fishpond to obtain the object space coordinate of the to-be-detected key point of the to-be-detected fishpond, and verification is carried out through the screened pointer point. According to the invention, approaching measurement of a plurality of large fishponds can be realized, high-precision measurement data can be obtained, and the measurement efficiency can be improved.

The application discloses a multi-angle coal gangue stock dynamic monitoring method based on multi-modal space-time perception, which comprises the following steps: constructing a multi-angle satellite coal gangue stock monitoring model containing a coal gangue recognition model; the coal gangue recognition model performs coal gangue recognition detection and constructs a mapping collinear equation for converting pixel coordinates into three-dimensional geographic coordinates; position point feature matching is performed on the cross-view features and space-time features of a multi-angle satellite image data set; multi-angle satellite image data set feature extraction, height perception coding processing and spectrum-guided cross-attention processing are performed to obtain multi-view stereoscopic images; homography projection is performed to a multi-view projection perspective plane, and detection and segmentation processing are performed respectively; then, coal gangue three-dimensional structures are fused and scaled to calculate the coal gangue volume. The application realizes high-precision and high-efficiency dynamic monitoring of the coal gangue volume and stock change by using multi-modal information such as spectrum, space-time and height.

The application relates to a method and device for monitoring an ancient road and a garden ecological dynamic based on binocular camera shooting, and belongs to the technical field of photogrammetry and computer vision. The method comprises the following steps: collecting multi-focus section images of the ancient road and the surrounding ecology; inversely calculating camera internal parameters of the corresponding focus section, and performing zoom camera endless zoom distortion calibration; using a collinear equation of binocular images and three-dimensional coordinates of control points to solve camera external orientation elements; taking a reference frame as a reference, calculating a relative offset of a target region in a continuous frame through sub-pixel accuracy least square matching, converting the ground resolution into a real ground deformation variable, and dynamically monitoring ancient road and ancient tree deformation variables; extracting ecological elements through image segmentation, analyzing coverage and texture changes based on a time sequence, and monitoring garden ecological changes. The application realizes joint monitoring of ancient road structure deformation, ancient tree posture change and surrounding ecological state, and improves the intelligent and fine level of cultural site protection.

The invention discloses a multi-angle coal gangue stock dynamic monitoring method based on multi-modal space-time perception, and the method comprises the steps: constructing a multi-angle satellite coal gangue stock monitoring model comprising a coal gangue recognition model, the coal gangue recognition model carries out coal gangue recognition detection and constructs a mapping collinear equation for converting pixel coordinates into three-dimensional geographic coordinates; position point feature matching is carried out by integrating cross-view features and spatial-temporal features of the multi-angle satellite image data set, and feature extraction, height perception coding processing and spectrum-guided cross attention processing are carried out on the multi-angle satellite image data set to obtain a multi-view three-dimensional image. Projecting to a multi-view projection perspective plane through homography, respectively carrying out detection and segmentation processing, fusing to form a three-dimensional structure of the coal gangue, and carrying out scaling calculation according to a proportion to obtain the volume of the coal gangue. According to the invention, high-precision and high-efficiency dynamic monitoring of the volume and stock change of the coal gangue is realized by using multi-modal information such as spectrum, space-time, height and the like.

This application relates to a method, apparatus, terminal device, and storage medium for determining camera relative pose parameter values ​​based on binocular vision. The method includes: obtaining a first relation by taking the partial derivative of each matrix element in a first camera extrinsic parameter matrix according to the collinearity equation; obtaining a second relation by taking the partial derivative of each relative pose parameter according to the relative pose parameter matrix and determining its product with a preset second camera extrinsic parameter matrix; determining a partial derivative function relation for the collinearity equation with respect to the relative pose parameters based on the product of the first and second relational expressions; obtaining a pixel deviation relation between the true image coordinates and the collinearity equation based on the product of the increment of the relative pose parameters and the partial derivative function relation; iterating over the pixel deviation relation, and obtaining the relative pose parameter values ​​of the first camera when the iteration is complete. This method can reduce computational complexity and improve camera calibration efficiency.

The invention relates to a component deformation determination method and device, computer equipment and a storage medium. Comprising the steps of obtaining point cloud data of a part and image data, collected by at least two image collection devices, of the part; aiming at each piece of image data, performing parallax correction on the image data according to the point cloud data, the first collinearity equation and external parameters of image acquisition equipment for acquiring the image data to obtain a correction graph corresponding to the image data; performing pixel matching on the correction images corresponding to different image data to obtain a correction pixel coordinate group corresponding to each position point of the component; therefore, parallax correction of the image data acquired by different acquisition devices is realized, the actual three-dimensional coordinate of each position point is determined according to each correction pixel coordinate group, the determined actual three-dimensional coordinate is more accurate, and according to the difference value between the actual three-dimensional coordinate of each position point and the initial three-dimensional coordinate, the actual three-dimensional coordinate of each position point can be corrected according to the difference value between the actual three-dimensional coordinate of each position point and the initial three-dimensional coordinate. And the deformation occurrence condition of the component is determined, so that the determined deformation occurrence condition is more reliable.

The application discloses a kind of building group regularizing method and system, system includes vector determination module, equation determination module and coordinate solution module.First, vector determination module obtains the vector polygon corresponding to building group, and the direction vector of building group is determined by vector polygon.Then, equation determination module identifies collinear contour edge and non-collinear contour edge in building group by direction vector, and determines collinear equation according to various parameters corresponding to all collinear contour edge, and determines straight line equation according to each parameter corresponding to each non-collinear contour edge.Finally, coordinate solution module can solve the building group contour node coordinates after regularizing with collinear equation and straight line equation as control condition, obtain the building group polygon after regularizing, realize building group regularizing, retain the topological relationship characteristics of adjacent contour edge at the same time, and avoid the emergence of gap and overlap etc.

The application relates to the technical field of image matching, and discloses a UAV image tilt angle anomaly detection method based on image matching and geometric model analysis, which comprises the following steps: acquiring a UAV image to be detected and a reference orthographic image; generating two-dimensional ground control points through feature matching and forming three-dimensional control points by supplementing elevations by using a public digital elevation model; solving photographic exterior orientation elements through monocular space resection; applying a collinearity equation model based on the solved exterior orientation elements to perform light ray tracing from an image side to an object side on image four-corner points, iteratively intersecting with a digital elevation model to accurately invert three-dimensional geographic coordinates of the ground four-corner points; constructing a geographic footprint quadrilateral based on the positions of the corner points and calculating a length ratio of upper and lower edges as a tilt angle index; and comparing the ratio with a preset threshold to determine a tilt angle anomaly. The application does not need to rely on an airborne attitude sensor at all, can invert a UAV photographic attitude based on image content, and can accurately compensate projection differences caused by terrain height differences through a projection geometric model under terrain constraints, so that high-precision automatic detection of UAV image tilt angle anomalies under complex terrain conditions is realized.

The present invention relates to the technical field of geological logging, and specifically to a three-dimensional space coordinate positioning method for coal-rock geological images, including: installing two cameras with fixed relative position information in parallel on a turntable, obtaining the camera parameters of the cameras, respectively taking overlapping geological images from both sides by the two cameras, and performing distortion correction on the geological images and resolving angular orientation elements according to the relative position information; performing gray processing on the geological images taken by the two cameras, extracting a preset number of homologous image points in the gray-processed geological images and collecting their pixel coordinates; constructing a collinearity equation set for each camera by using the space resection method according to the camera parameters and pixel coordinates, and calculating the interior orientation elements, exterior orientation elements, and distortion residuals. The present invention only needs to extract a set of data for calculation, reduces the amount of calculation, and overcomes the errors caused by the inability to use cameras with strictly collinear optical axis centers, improving the accuracy and efficiency of photogrammetry.

The invention provides a method and a device for determining an aviation large-inclination-angle photographing route, which are applied to the technical field of aviation remote sensing. Comprising the following steps: converting the longitude and latitude of a target area into projection plane coordinates, constructing a bounding rectangle surrounding a target, and calculating an initial side-looking sweep angle by taking a long side as a flight direction and combining a minimum approaching distance, a flight height and a rectangle parameter; determining the coverage width of a measurement area according to the rectangular width and the lateral extension proportion, calculating the swing-scan field angle, the cycle index and the interval angle corresponding to the maximum cut-off swing-scan angle, determining the Y coordinate of a route according to the rectangular coordinate and the minimum approaching distance, and calculating the X coordinate of an exposure point to form a sequence in combination with the camera course breadth and the overlapping degree. And calculating the position attitude of the aerial photo based on the exposure point serial number, the sweep angle and the route direction, calculating the ground coordinates and elevation of four corners of the aerial photo through a collinear equation by combining camera parameters and a digital elevation model, and determining the coverage range.

The application relates to a progressive adjustment continuous pose estimation method for a lunar probe landing process, which comprises the following steps: step 1) selecting a landing camera sequence image from a rapid adjustment section to a hovering section, obtaining multiple connection point matching information through feature matching, and performing initial pose estimation by adopting a 'double model relative orientation-three image model connection'; step 2) performing pose parameter refinement on the initial pose information by using a sliding window type coplanar or collinear equation constraint, combining Schur complement edge linearization and weighted pseudo-observation constraint, and step 3) constructing an incremental collinear adjustment model of additional external control points to optimize the result of step 2), so as to reduce the pose estimation and positioning error. Compared with the prior art, the application has the advantages of effectively realizing high-precision and rapid calculation of continuous position and attitude parameters in the lunar probe landing process, and the incremental collinear adjustment model of additional external control points can further improve the pose estimation and positioning precision.

The invention provides a water surface-oriented unmanned aerial vehicle image ortho-rectification method and device, which are applied to the technical field of unmanned aerial vehicle surveying and mapping, and the method comprises the following steps: obtaining an unmanned aerial vehicle image obtained by shooting a water surface area by an unmanned aerial vehicle; coordinate system conversion is carried out based on the longitude and latitude of the photography center, and plane coordinates of the photography center in a photography measurement coordinate system are obtained; determining a ground point coordinate of each pixel in the unmanned aerial vehicle image in the photogrammetry coordinate system according to a collinear equation based on the plane coordinate of the photogrammetry center and the unified water surface elevation value; constructing point cloud data of the unmanned aerial vehicle image based on the ground point coordinate of each pixel; determining a ground sampling distance based on the geodetic height of the unmanned aerial vehicle image, the unified water surface elevation value and the equivalent focal length of the unmanned aerial vehicle during shooting; and based on the ground sampling distance, carrying out orthographic rasterization processing on the point cloud data of the unmanned aerial vehicle image to obtain an image after orthographic correction. According to the invention, the defect of excessive dependence on ground feature points in the prior art is overcome.

The present invention relates to a method for calculating the railway mileage of an unmanned aerial vehicle based on deep learning, including: obtaining geographic information data; training a deep learning model; performing frame-by-frame detection and segmentation on a video; calculating ground coordinates using the collinearity equation; calculating relative displacement through calculation points; and post-processing the mileage. The present invention overcomes the disadvantages of traditional manual inspections, which are completed on foot for a long time, with low work efficiency, untimely information upload, difficult effective supervision of operations, and requiring a large amount of time cost, reducing the maintenance cost of the line; provides line mileage information, avoids the problem of full-section manual inspections when foreign objects appear on the line, and speeds up the time for line maintenance personnel to arrive at the scene for handling; using an unmanned aerial vehicle for inspection has no requirement for skylight time, can perform flight detection for a long time, and can timely discover problems and handle or report them.

The application provides a kind of unmanned aerial vehicle image ortho rectification method and device for water surface, applied to unmanned aerial vehicle surveying and mapping technical field, above-mentioned method includes: obtaining unmanned aerial vehicle image that unmanned aerial vehicle is shot to water surface area;Based on the longitude and latitude of photographic center, coordinate system conversion is carried out, and the plane coordinates of photographic center in photogrammetry coordinate system are obtained;According to collinear equation, based on the plane coordinates of photographic center and unified water surface elevation value, the ground point coordinates of each pixel in unmanned aerial vehicle image in photogrammetry coordinate system are determined;And based on the ground point coordinates of each pixel, the point cloud data of unmanned aerial vehicle image is constructed;Based on the geodetic height of unmanned aerial vehicle image, unified water surface elevation value and equivalent focal length when unmanned aerial vehicle is shot, ground sampling distance is determined;Based on ground sampling distance, the point cloud data of unmanned aerial vehicle image is ortho rasterization processing, and the image after ortho rectification is obtained.Through the present application, the defect that over-reliance on ground feature points in related art is solved.

The invention relates to the technical field of vision measurement, and provides a large-view-field vision measurement method and a large-view-field vision measurement system.The method comprises the following steps that S1, a calibration method of vision cameras applied to large-view-field optical measurement is adopted to calibrate each vision camera, and internal parameters and external parameters of each vision camera are obtained; s2, simultaneously shooting a target to be detected by the visual camera, and obtaining a shot image; s3, obtaining an image point of the to-be-measured target from the shot image; s4, simultaneous collinear equations are established; and S5, solving the collinear equation by adopting a least square method according to the internal and external parameters and the image points of the visual camera to obtain the position information of the to-be-measured target. According to the scheme, simple and high-precision calibration can be carried out before measurement, the point distribution difficulty is reduced, and the measurement precision in a view field is ensured.