KR102127679B1 - System for correcting geometry of mobile platform with sensor based on an orthophoto - Google Patents

Abstract

A geometric correction system for a mobile platform equipped with a sensor based on orthoimage is provided.
The system includes a sensor unit including at least one of an image sensor, a navigation sensor, and a three-dimensional surveying sensor that acquires three-dimensional geographic data, and observation-related data and two-dimensional terrain obtained from at least one sensor mounted on the mobile platform. Based on a similarity analysis unit that performs similarity analysis between orthogonal image-related data having information and sets reference data that is equally estimated from the observation-related data and the orthoimage-related data, and the reference data and the reference coordinate system of the orthoimage The network model constructing unit for generating a geometric model between the reference data and the observation-related data, and calculates a most probable value that minimizes the amount of separation between the observation-related data and the reference data in the geometric model. And a calculation unit for calculating correction data estimated from the observation-related data according to the determination value.

Description

System for correcting geometry of mobile platform with sensor based on an orthophoto

The present invention relates to a geometric correction system for a mobile platform equipped with a sensor based on a normal image, and more specifically, to analyze the similarity between the observation related data and the normal image of the sensor mounted on the mobile platform by utilizing the feature geometry in the normal image. And, by generating a geometric model between the two data based on the reference data thereby, it relates to a system for securing the accuracy and stability of the three-dimensional geometric information correction of the mobile platform using an orthogonal image composed of two-dimensional terrain information.

Mobile mapping and autonomous driving systems that are under development are equipped with lidars, cameras, GPS/navigation sensors, and are used for situational awareness and HD map generation. These sensors may be mounted on a mobile platform for mobile mapping for map production or an autonomous mobile platform for recognition of surrounding geographical conditions during autonomous driving. In order to improve the accuracy and stable operation of the mobile platform, the accuracy of location information must be sufficiently secured.

Advances in GNSS (Global Navigation Satellite System) and INS (Inertia Navigation System) have contributed significantly to the improvement of accuracy (50cm level), but the accuracy is sufficient to be used in the stable autonomous driving system of mobile platforms and the mapping for this. Level).

As the placement of sensors changes during the operation of a mobile platform, the internal model of the platform needs to be calibrated through monitoring at any time. For this reason, correction of the geometry of the mobile platform is required, and this is generally done by acquiring and correcting the ground reference point through GNSS' Network-RTK survey.

For example, even in manufacturing with precision, expensive MMS equipment equipped with a laser sensor and an image sensor is used, but due to the limitation of geometric accuracy, the process of separately performing a reference point survey through GNSS survey is essential for quality calibration It is included as a process.

In recent drone surveying, reference surveying through GNSS surveying is included as an essential process, but it is difficult to secure a reference point when shooting difficult access areas such as mountainous areas and waterside spaces.

However, in the case of GNSS survey, it requires a professional survey that requires a lot of time and manpower, which incurs a great deal of cost, so it is a big obstacle to take advantage of the mobile platform.

In the related art, since the accuracy of the position of the mobile platform and the stability of the geometric model are not sufficiently guaranteed, it is necessary to perform the GNSS survey to increase the accuracy. However, this also has the disadvantage of high cost/manpower/time consumption and difficult measurement in difficult access areas.

In order to overcome this, a method of adopting a method of acquiring and applying a reference point from an already built satellite image or an orthogonal image produced by an aerial image system has been attempted. The orthoimage is image information reflecting an orthogonal projection model in which distortion such as undulation displacement caused by the projection geometric model of the camera is removed, and is data of two-dimensional topographic information mainly used for digital map production.

Since ortho-images are data managed by a national institution for public purposes, the criteria for quality control for location accuracy and resolution are clear, and there is an advantage of easy quality control for utilization results.

However, some problems are caused when using orthoimages.

Since the orthographic image-related data is large-capacity, a lot of time is devoted to searching for data that matches the mobile platform data to be corrected. In addition, since the spatial information type or characteristic between the orthoimage and the acquired mobile platform data is different, the accuracy of the similarity analysis is not sufficiently secured. In addition, since the orthogonal image is generally composed of two-dimensional coordinates, there is a limit to apply it to three-dimensional geometric correction for a mobile platform.

The technical problem to be achieved by the present invention is to analyze the similarity between the observation-related data and the orthogonal image of the sensor mounted on the mobile platform by utilizing the feature geometry in the orthogonal image, and based on this, the geometric model between both data By creating, it is to provide a geometric correction system of a mobile platform equipped with a sensor that ensures accuracy and stability of 3D geometric information correction of a mobile platform using an orthogonal image composed of 2D terrain information.

The objects of the present invention are not limited to the objects mentioned above, and other objects not mentioned will be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention for achieving the above technical problem, a geometric correction system of a mobile platform equipped with a sensor based on a normal image includes at least one of an image sensor, a navigation sensor, and a three-dimensional survey sensor for obtaining three-dimensional geographic data. Similarity analysis is performed between the observation unit and the orthogonal image-related data by performing a similarity analysis between the sensor unit and observation-related data obtained from at least one sensor mounted on the mobile platform and orthoimage-related data having two-dimensional terrain information. A similarity analysis unit for setting the estimated reference data, a network model construction unit for generating a geometric model between the reference data and the observation-related data based on the reference data and the reference coordinate system of the orthogonal image, and the geometric model And a most probable value for minimizing the amount of separation between the observation-related data and the reference data, and includes a most-value calculation unit for generating correction data estimated from the observation-related data according to the maximum value.

In another embodiment, the orthoimage is image information reflecting an orthogonal projection model in which distortion caused by the projection geometric model of the camera is removed, and the observation-related data is observation data directly output from each sensor and the position of the mobile platform And external geometry that defines the geometric relationship of the sensor according to the posture, time information for the acquisition time of each sensor, point group data related information based on the 3D survey sensor, geocoding image data related information based on the image sensor , Based on at least one of the object information, the attribute of the extracted object based on at least one of the point group data related information and the geocoding image data related information, the point group data related information, and geocoding image data related information. , May include at least one of map-related information generated to create a map.

Here, the observation-related data further includes an internal geometry calculated based on geometric parameters uniquely defined for each sensor, and the network model constructing unit is the reference data based on geometric information composed of the external geometry and the internal geometry. And a geometric model between the observation-related data.

In addition, the maximum value is the geometry information generated by at least one of the external geometry, the error information of the internal geometry, the point group data related information, the geocoding image data related information, the object information, and the map related information. At least one selected from error information, and the geometric shape information may be information related to feature geometry estimated to be the same from the reference data and the observation related data.

In another embodiment, the orthogonal image search unit for retrieving and extracting orthogonal data for a region of interest that matches the movement trajectory of the mobile platform from the orthogonal image further includes the orthogonal image-related data in a feature geometry within the region of interest And orthogonal data having geometric shape information estimated to be, and the geometric shape information may be extracted based on a specific geometric shape of the object extracted in the region of interest.

Here, the similarity analysis is performed to set the reference data that is estimated equally between the orthonormal data and the observation related data, and when the similarity analysis unit acquires the observation related data from the image sensor, the similarity analysis is performed on the observation related data. Performs similarity analysis to set the reference data by analyzing color change based on pixel coordinates among the geocoding image data related information, and obtains the observation related data from the 3D survey sensor In the case of analysis, similarity analysis may be performed by extracting the geometric shape information and setting the reference data by analyzing at least one of the geometric shape change degree and the laser intensity change degree of the point group data belonging to the observation-related data.

In another embodiment, before the similarity analysis is executed, the similarity analysis unit may perform pre-processing of image luminance adjustment, pre-processing of image resolution, pre-processing of image clarity, and pre-processing of image clarity in at least one of the observation-related data and the ortho-image-related data. At least one pre-processing selected from pre-processing for transformation, pre-processing for transforming an image coordinate system, and pre-processing for extracting boundary information of an object from the observation and orthogonal image-related data may be performed.

In another embodiment, predicted error probability analysis for determining whether the expected error exceeds the allowable error by analyzing the expected error between the correction data and the observation-related data based on the minimum separation amount used when calculating the maximum value Further comprising, the predicted error probability analysis unit, when the expected error exceeds the allowable error, the process of retrieving and extracting the orthogonal image-related data by adjusting the region of interest corresponding to the movement trajectory of the mobile platform, When the similarity analysis unit performs pre-processing to minimize the amount of deformation due to a difference in observed geometric characteristics between the observation-related data and the ortho data before the similarity analysis, the conditions of the pre-processing are changed, or the ortho-image related data and the observation-related A process of sampling and extracting geometric shape information estimated to be another object or other feature geometry from data as new reference data, and processing to remove observation-related data determined as abnormal data from the geometric model generated by the network model construction unit , A process of changing at least a part of parameters applied to the geometric model, and a process of changing at least some parameters used in the analysis model equation of the most applicable value applied to the most accurate value calculation unit or changing the analysis model equation The similarity analysis unit, the network model construction unit, and the maximum value calculation unit may be controlled to perform at least one of the above, and may be controlled to generate the maximum value and correction data again according to constraints after performing the processing.

In another embodiment, the network model construction unit further comprises a supplementary process of adding vertical geometric data to the reference data applied to the geometric model, or the reference without vertical geometry data when constructing the geometric model Data can be applied.

In addition to this, the complementary processing is a process of setting the vertical geometric data to a predetermined value, a process of inverting three-dimensional coordinates with a plurality of reference data, the vertical representation of a raster-shaped digital elevation model or a vector-shaped numerical map. The observation range for each of the plurality of points from a multi-view or stereo-view in which the sensor is acquired at multiple points according to the process of adding geometric data to the reference data and the movement trajectory of the mobile platform It may be performed by at least one of the process of calculating the vertical geometric data by configuring the intersection geometry of the regions overlapping each other.

Specific details of other embodiments are included in the detailed description and drawings.

According to the present invention, the present invention relates to a geometric correction system of a mobile platform equipped with a sensor based on a normal image, and more specifically, analysis of the similarity between the observation related data and the normal image of the sensor mounted on the mobile platform using the feature geometry in the normal image By performing the, and generating a geometric model between the two data based on the reference data thereby, it is possible to secure the accuracy and stability of the three-dimensional geometric information correction of the mobile platform using an orthogonal image composed of two-dimensional terrain information.

FIG. 1 is a diagram illustrating a process in which sensors mounted on various mobile platforms in the region of an orthoimage acquire data related to observation.
2 is a configuration diagram of a geometric correction system of a mobile platform equipped with a sensor based on a normal image according to an embodiment of the present invention.
3 is a flowchart illustrating a process of estimating correction data implemented in a geometric correction system of a mobile platform equipped with a sensor.
4 is a view for explaining an orthogonal image distinguished from a general camera image.
FIG. 5 is a diagram illustrating a process of detecting and extracting orthogonal data in a region of interest according to the movement trajectory of the mobile platform in the orthogonal image.
FIG. 6 is a diagram illustrating a process of extracting geometric shape information from orthoimages and observation-related data in a similarity analysis unit.
7 is a diagram illustrating a process of setting reference data based on geometric shape information extracted from an orthoimage and observation-related data in a similarity analysis unit.
FIG. 8 is a diagram illustrating a process of deriving correction data from the most accurate value calculation unit while generating a geometric model between reference data and observation related data in the network model construction unit.
FIG. 9 is a diagram illustrating processing for calculating 3D geometric data by constructing an intersection geometry from a multi-view or stereo-view according to a movement trajectory of a mobile platform.
10 is a flowchart illustrating a processing process when an expected error according to a maximum value exceeds an allowable error in a geometric correction system of a mobile platform equipped with a sensor according to an embodiment of the present invention.
FIG. 11 is a diagram in which correction data geometrically corrected based on a normal image is observed in observation-related data taken by a 3D survey sensor.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings and the contents described below. However, the present invention is not limited to the embodiments described herein and may be embodied in other forms. Rather, the embodiments introduced herein are provided to ensure that the disclosed contents are thorough and complete and that the spirit of the present invention is sufficiently conveyed to those skilled in the art. Throughout the specification, the same reference numbers refer to the same components. On the other hand, the terms used in this specification are for describing the embodiments and are not intended to limit the present invention. In the present specification, the singular form also includes the plural form unless otherwise specified in the phrase. As used herein, "comprises" and/or "comprising" refers to the presence of one or more other components, steps, operations and/or elements in which the mentioned component, step, operation and/or element is present. Or do not exclude additions.

In addition, "parts" to modules generally refer to parts such as software (computer programs) and hardware that are logically separable. Therefore, the module in this embodiment indicates not only a module in a computer program, but also a module in a hardware configuration. Therefore, the present embodiment provides a computer or program for functioning as a module (a program for executing each step on the computer, a program for functioning the computer as a respective means, and a function for realizing each function on the computer). Program), system and method. However, for convenience of explanation, "save", "save", and equivalents to these are used, however, these sentences are stored in a storage device or stored in a storage device when the embodiment is a computer program. It means to control like this. Further, the "parts" to the modules may correspond to the functions one-to-one, but in implementation, one module may consist of one program, multiple modules may consist of one program, or conversely, one module may consist of multiple programs. do. Further, multiple modules may be executed by one computer, or one module may be executed by multiple computers by a computer in a distributed or parallel environment. In addition, another module may be included in one module. In addition, hereinafter, the "connection" is also employed in the case of a logical connection (transmission and reception of data, instructions, reference relations between data, etc.) in addition to a physical connection. The term "predetermined" means that it is determined before the targeted treatment, and not only before the treatment according to the present embodiment is started, but also after the treatment according to the present embodiment is initiated, before the targeted treatment is performed. It is used by including the meaning of what is determined according to the situation, condition, or situation until then.

In addition, a system or device is configured by connecting a plurality of computers, hardware, devices, and the like by a communication means such as a network (including one-to-one communication connections), and by using one computer, hardware, device, or the like. It is also included when it is realized. The terms "apparatus" and "system" are used as synonymous terms. Of course, the "system" does not include anything other than a social "organization" (social system), which is an artificial decision.

In addition, when processing is performed by each unit or each module or when a plurality of processing is performed within each unit or module, the target information is read and input from the storage device, and after processing, the result of processing is processed. It is to write in the storage device. Therefore, description may be omitted regarding read input from the storage device before processing and writing to the storage device after processing. Further, the storage device may include a hard disk, random access memory (RAM), an external storage medium, a storage device via a communication line, a register in a CPU (Central Processing Unit), and the like.

Hereinafter, with reference to FIGS. 1 and 2, a geometric correction system of a mobile platform equipped with a sensor based on an orthogonal image according to an embodiment of the present invention will be described in detail.

1 is a view showing a process of acquiring observation-related data by sensors mounted on various mobile platforms within an area of a normal image, and FIG. 2 is a mobile platform equipped with a sensor based on a normal image according to an embodiment of the present invention. This is a schematic diagram of a geometric correction system.

The geometric correction system 100 of the sensor-mounted mobile platform is mounted on a platform such as a ground and/or air vehicle, and may be used as a mobile mapping system or a system for autonomous driving depending on the application.

For example, as shown in FIG. 1, the mobile platform equipped with a sensor may be an unmanned aerial vehicle 14 such as a vehicle 10, an aviation 12, or a drone, and the sensor mounted on the mobile platform may be viewed in red. Observe multiple sensor data while observing.

Referring to FIG. 2, the geometric correction system 100 of the mobile platform equipped with a sensor includes an image sensor 102, a three-dimensional survey sensor 104, a navigation sensor 106, an orthogonal image search unit 108, and a sensor data processing unit ( 110), a similarity analysis unit 112, a network model construction unit 114, a most probable value calculation unit 116, and a predicted error probability analysis unit 118. The image sensor 102, the three-dimensional survey sensor 104, and the navigation sensor 106 constitute a sensor unit, and in this embodiment, it is mainly described that three heterogeneous sensors are mounted, but the sensor unit is at least one of the sensors described above. Examples including the same are not excluded.

The image sensor 102 is a sensor that is mounted on a ground or an aerial mobile platform and acquires observation data for an image by photographing surrounding objects, such as terrain and features, as an image. And is not limited thereto.

The 3D survey sensor 104 is a sensor that is mounted on a platform and acquires 3D geographic data related to objects around it, such as terrain and features, and acquires observation data for 3D survey, and is an active remote sensing sensor. . For example, the 3D measurement sensor 104 may be a laser or an ultrasonic sensor, and in the case of a laser sensor, it may be a light detection and ranging (LiDAR) sensor. The lidar sensor scans a laser to an object to acquire data and detects a parallax and energy change of an electromagnetic wave reflected and returned from the object to calculate a distance and reflection intensity for the object.

The navigation sensor 106 is a sensor that detects navigation information such as positioning information, platform position, posture, and speed, and acquires observation data for navigation, and acquires a position for acquiring a moving position of the platform through a satellite navigation device (GPS) The apparatus may include a device, an inertial measurement unit (IMU), and a posture acquisition device that acquires a vehicle's posture through an inertial navigation system (INS).

Each sensor may be mounted on the same platform, but may be distributed and mounted on a ground platform and an aerial platform such as satellite, aviation, and drone.

As shown in FIG. 5, the orthogonal image search unit 108 receives a plurality of orthogonal image-related data for a region of interest corresponding to the movement trajectory 18 of the mobile platform from the orthogonal image 16 shown in the left image of FIG. 5. You can search and extract.

The orthogonal image 16 is image information reflecting an orthogonal projection model in which distortion such as undulation displacement caused by the projection geometric model of the camera is removed, and is data of two-dimensional topographical information mainly used for digital map production. Specifically, as shown in FIG. 4, camera image information obtained from a camera or a navigation sensor of a satellite or aeronautical system basically has a perspective projection geometry. An image that is modeled to have the geometric characteristics of orthogonal projection through camera correction and geometric modeling (Back-projection, 3D intersection, etc.), etc., is an orthogonal image, and is horizontally positioned on each pixel of the orthogonal image. The coordinates are matched. Orthoimages obtained from satellite images or aerial images are constructed all over the world and can be updated periodically according to the shooting cycle. Orthoimages are managed by the National Geographic Information Institute in Korea, and since the standards for quality control for location accuracy and resolution are clear, quality control for utilization results is very easy.

The orthoimage-related data includes orthogonal data having geometric shape information estimated as feature geometry in the region of interest, and the geometric shape information may be extracted based on a specific geometric shape of the object extracted in the region of interest. The object may be randomly selected as an object capable of easily extracting feature geometries such as points, lines, faces, and columns on a map, such as a road sign, an object of unusual shape, and a building boundary line.

As in the present embodiment, by setting the region of interest corresponding to the movement trajectory 18 of the mobile platform to search for and extract ortho data, contrast with observation-related data processed from sensor data obtained from the sensors 102-106 Not only can the time required to set the reference data and analyze the similarity be shortened, but also the accuracy in the process of estimating the correction data can be improved.

The sensor data processing unit 110 may output a plurality of observation-related data by performing predetermined processing on sensor data obtained from the at least one sensor 102-106.

Observation-related data is calculated based on observation data output directly from each sensor (102~106), external geometry defining the geometric relationship of the sensor according to the position and attitude of the mobile platform, and geometric parameters uniquely defined for each sensor. It may include at least one of the internal geometry. The external geometry and the internal geometry can constitute geometric information.

In addition, the observation-related data includes the above-described observation data, geometric information, time information for the acquisition time of each sensor, unique data related to device-specific attributes for each sensor 102-106, point group based on the three-dimensional survey sensor 104 Based on at least one of data-related information, geocoding image data-related information based on the image sensor 102, point group data-related information, and geocoding image data-related information, object information, point group estimated by attribute of the extracted object Based on at least one of data-related information and geocoding image data-related information, at least one of map-related information generated to create a map may be included.

The internal geometry is calculated based on geometric parameters defined for each sensor 102-106, and in this case, may be calculated through a predetermined equation.

The internal geometry is an inherent value of the sensor itself, and is an error in observation data for each sensor 102 to 106 due to parameters maintained regardless of whether the platform or the like is moved.

Here, the geometric parameter for the image sensor may be at least one of a focal length, a pub position, a lens distortion parameter, and a sensor format size. For example, the geometric parameter for the 3D survey sensor may be at least one of an incident angle, a distance scale, a distance offset, and an axial offset of each laser, for example. Further, the geometric parameter for the navigation sensor may be at least one of an axial scale and an axial offset.

The detailed calculation process of the internal geometry performed by the sensor data processing unit 110 will be described later.

The sensor data processing unit 110 is an external geometry that defines an internal geometry for each sensor 102-106 and a geometric relationship between each sensor according to a position and posture of a platform on which at least one of the sensors 102-106 is mounted. Data fusion is performed by performing geometric modeling processing based on the configured geometric information.

The external geometry is calculated based on the parameter values that change each time the observation data of each sensor 102-106 is acquired due to the position and attitude of the platform being moved, that is, the position and attitude of each sensor 102-106. It is an error of the observed data, and in this case, it can be calculated through a predetermined equation.

The detailed calculation process of the external geometry performed by the sensor data processing unit 110 will be described later.

In addition to this, for example, for each sensor 102 to 106 for the unique data, the unique data of the image sensor 102 includes sensor illumination, ISO, shutter speed, shooting date/time, time synchronization information, sensor model , Lens model, sensor serial number, image file name, and file storage location.

The unique data of the 3D survey sensor 104 includes, for example, lidar information, the date and time of the sensor, sensor model, sensor serial number, laser wavelength, laser intensity, laser transmission/reception time, laser observation angle/distance, laser pulse , Electronic time delay, standby delay, and sensor temperature.

The unique data of the navigation sensor 106 includes GNSS/IMU/INS sensor model information, GNSS reception information, GNSS satellite information, GNSS signal information, GNSS navigation information, ionospheric/convective signal delay information of the GNSS signal, DOP information, and Earth. Behavior information, dual/multiple GNSS equipment information, GNSS base station information, wheel sensor information, gyro sensor scale/bias information, accelerometer scale/bias information, position/posture/velocity/acceleration/angular velocity/angular acceleration information and expected error, navigation It may include at least one of information filtering model and erasure error information, photographing date/time, time synchronization information, or all.

The time information may be a time when observation data of each sensor 102-106 is generated.

The sensor data processing unit 110 may generate 3D point cloud data related information based on the 3D survey sensor 104 and geocoding image data related information based on the image sensor 102 based on the observation data and geometric information. . In this embodiment, an example of generating all of the point group data and geocoding image data related information is provided, but in some cases, only some of them may be generated.

The 3D point cloud data-related information includes point group 3D coordinate data, point group color information, point group class information estimated from the object extracted from point group 3D coordinate data, and laser when the 3D survey sensor is a laser sensor. It may include at least one of the intensity information.

The geocoding image data-related information may include at least one of geocoding coordinate data for each pixel, color information for geocoding, and class information for geocoding whose type is estimated from an object extracted from the geocoding coordinate data. .

The sensor data processing unit 110 extracts object information related to the attribute of the extracted object based on at least one of the 3D point cloud data related information and the geocoded image data related information, and also extracts map related information for creating a map. You can create and map it through the drawing process.

The sensor data processing unit 110 may recognize information on the object by applying machine learning to candidate group data related to the object identified from at least one of the 3D point group data related information and the geocoding image data related information. In this case, the object information may include properties related to a specific geometric shape, color, and texture, such as an object type, an overall shape, and lines, faces, circles, and spheres constituting at least a part of the object.

The map-related information may include at least one of map coordinate data based on at least one of point group 3D coordinate data and geocoding coordinate data, and object information estimated from an object.

In addition, the map-related information, the 3D coordinate data included in the 3D point cloud data related information, and the geocoding coordinate data included in the geocoding image data related information are two-dimensional or three-dimensional transformation data, such as Rigid-body, Affine, Similartiy It may be composed of any one of the data.

The similarity analysis unit 112 performs predetermined pre-processing on at least one of observation-related data and orthogonal data obtained from at least one sensor 102 to 106 mounted on a mobile platform and processed by the sensor data processing unit 110. , Similarity analysis between observation-related data and orthogonal data may be performed to set at least one reference data that is equally estimated from multiple observation-related data and multiple orthonormal data.

The above-described pre-processing is at least one of pre-processing of image luminance adjustment, pre-processing of image resolution adjustment, pre-processing of image sharpness adjustment, pre-processing of image geometric transformation, pre-processing of image coordinate system transformation, and pre-extraction of object boundary information from observation and ortho-image related data. Can be applied. This pre-processing improves the accuracy of similarity analysis by minimizing the amount of deformation such as geometric information, coordinate system, and resolution due to differences in the observed data characteristics such as the unique data of the sensors 102 to 106, the environment during observation, and scale/rotation/displacement values. To be carried out. For example, an orthogonal image of orthogonal data is acquired in the clear weather at noon, and when observation-related data is acquired during rainy weather at sunset, the resolution of both data may be different and errors in similarity analysis may be caused. In order to solve this, pre-processing to adjust the resolution of both data to a close value may be performed. The image geometric transformation pre-processing may be image scale transformation or image rotation change.

In addition, when obtaining observation-related data from the image sensor 102, the similarity analysis unit 112 analyzes color changes based on pixel coordinates among geocoding image data-related information belonging to the observation-related data to obtain geometric shape information. Similarity analysis can be performed by extracting and setting reference data. When acquiring observation-related data from the three-dimensional survey sensor 104, the similarity analysis unit 112 analyzes at least one of the geometric shape gradient and the laser intensity gradient of the point cloud data belonging to the observation-related data. Similarity analysis can be performed by extracting information and setting reference data.

According to this, it is possible to improve the matching precision by analyzing the similarity between the reference data and the object information of the observation-related data, for example, geometrical information such as the type of object or specific geometric shape.

The network model construction unit 114 generates a geometric model between the reference data and the observation-related data based on the reference data and the reference coordinate system of the orthoimage.

Specifically, the network model construction unit 114 may generate a geometric model between reference data and observation related data based on geometric information composed of external geometry and internal geometry. In addition, the reference coordinate system may be set as the orthogonal image coordinate system 20 of the orthogonal image 16 shown in FIG. 1. By using an orthogonal image coordinate system that is the basis for shape modeling or numerical analysis, observation-related data can be converted into a standardized coordinate system and corrected.

As another embodiment, the network model construction unit 114 may generate a geometric model between reference data and collected data based on external geometry among geometric information. Considering that the positions and postures of the mobile platform on which the sensors 102 to 106 that acquire observation-related data are mounted and the sensors that acquire the orthoimage are different from each other, external geometry may be preferentially used when generating the geometric model.

The network model constructing unit 114 may calculate an objective function that minimizes the distance from observation-related data through a geometric model to which a plurality of parameters are applied.

The creation of the detailed geometric model will be described later with reference to FIGS. 3 and 8.

The network model constructing unit 114 may apply the reference data without vertical geometric data when constructing the geometric model, but the orthonormal data is two-dimensional data, and observation-related data obtained from the three-dimensional survey sensor 104 is 3 In the case of dimensional data, in order to increase the consistency of the geometric model, a supplementary process of adding vertical geometric data to the reference data applied to the geometric model may be further included.

In this case, the complementary processing is a process of setting vertical geometric data to a predetermined value, a process of inverting three-dimensional coordinates with a large number of reference data, a vertical geometry data in a raster form, a high-resolution model or a vector form, It may be at least one of the process of adding to the reference data and the process of adding three-dimensional geometric data by the intersection geometry 30 as shown in FIG. 9. The processing of FIG. 9 is a plurality of multi-view or stereo-view from which the sensors 102 to 106 are acquired at multiple points, as in FIG. 9 according to the movement trajectory 18 of the mobile platform. The vertical geometry data is calculated by constructing the intersection geometry 30 with respect to the feature geometry 22 of the reference data photographed in the regions overlapping each other in the observation range 28 for each point. This is a process that is added to the reference data 24 of the orthogonal image 24. In the case where the geometric stability is not secured by the complementary processing by the intersection geometry, the other processing described above may be additionally applied.

When observation-related data can be obtained in the form of a navigation sensor 106 combined with an image sensor 102, equidistant, equisolid-angle, orthographic, which are used for photo analysis, Stereoscopic, perspective geometry, longitudinal cylindrical, transverse cylindrical, mercarto, trasverse mercarto, spherical, orthographic, fisheye lens type, stereoscopic , Geometric modeling is possible with at least one of the central perspectives. If the observation-related data can be obtained in the form of a combination of the navigation sensor 106 and the three-dimensional survey sensor 104, geometric modeling is possible as a model including three-dimensional transformation geometry.

The most probable value calculator 116 calculates a most probable value that minimizes the amount of separation between observation-related data and the reference data in a geometric model, and generates correction data estimated from the observation-related data according to the most probable value. . The most accurate value calculating unit 116 uses a predetermined analysis model formula, and some parameters among a plurality of parameters may be applied to the model formula.

By estimating the correction data with the calculated maximum value, geometric correction of observation-related data using the reference data is performed, and such time-limit correction is the result of applying the reference data to the geometric model and the geometry of observation-related data corresponding to the reference data This is the process of analyzing the error amount and calculating the minimum or optimized parameter as the maximum value.

The maximum value may be at least one of external geometric and internal geometric error information. In addition, the maximum value may be calculated as at least one of error information of geometric shape information generated by at least one of the above-described information, point group data related information, geocoding image data related information, object information, and map related information. The process of calculating the maximum value will be described later with reference to FIG. 3.

In addition, the predicted error probability analysis unit 118 analyzes the predicted error between the correction data and the observation-related data based on the minimum separation amount used when calculating the maximum value, and determines whether the predicted error exceeds the allowable error, In case of exceeding, at least one of the following processes is instructed to the relevant part to be processed. The predicted error probability analysis unit 118 restarts the correction process from the corresponding unit (module) related to the newly set constraint condition after such processing, and controls to reproduce the maximum value and correction data.

The predicted error probability analysis unit 118 searches for orthoimages to perform a process of searching for and extracting orthonormal data by adjusting a region of interest that matches the moving trajectory 18 of the mobile platform when the predicted error exceeds the allowable error. The unit 108 can be controlled.

The predicted error probability analysis unit 118 changes at least one of the pre-processing performed by the similarity analysis unit 112, or uses the geometric shape information estimated as another object or other feature geometry in orthogonal data and observation-related data as a new criterion. The similarity analysis unit 112 may be controlled to perform a process of sampling and extracting data.

The predicted error probability analysis unit 118 may control the network model construction unit 114 to perform a process of determining and removing observation-related data representing a predetermined distance from the objective function calculated through the geometric model as abnormal data. .

The predicted error probability analysis unit 118 may control the network model construction unit 114 to perform a process of changing at least some parameters applied to the geometric model.

The predicted error probability analysis unit 118 is configured to perform at least one of the process of changing at least some parameters used in the analysis model formula of the most applicable value applied to the most accurate value calculation unit 116 or changing the analysis model formula. It is also possible to control the value calculation unit 116.

By the predicted error probability analysis unit 118, sufficient accuracy of the correction data can be further improved.

According to this embodiment, by utilizing the feature geometry in a two-dimensional orthogonal image, a similarity analysis between various techniques between observation-related data and an orthogonal image of a sensor mounted on a mobile platform is performed, and based on this, between the two data By generating a geometric model, it is possible to secure accuracy and stability of 3D geometric information correction of a mobile platform using an orthogonal image composed of 2D terrain information.

According to the present embodiment, the orthoimage is easy to secure and utilize the reference data in a one-to-one relationship with the coordinates of the ground, and can facilitate management of the reference data when performing geometric correction of the mobile platform. Based on this, it is possible to generate highly accurate correction data in correction of observation related data.

Hereinafter, a process of estimating correction data implemented in a system according to an embodiment of the present invention will be described with reference to FIGS. 1 to 9 and 11.

3 is a flowchart illustrating a process of estimating correction data implemented in a geometric correction system of a mobile platform equipped with a sensor, and FIG. 4 is a diagram illustrating an orthogonal image distinguished from a general camera image. 5 is a view showing a process of detecting and extracting ortho data from a region of interest according to a movement trajectory of a mobile platform from an ortho image, and FIG. 6 is a process of extracting geometric shape information from ortho images and observation-related data in a similarity analysis unit It is a figure showing. FIG. 7 is a diagram illustrating a process of setting reference data based on geometric shape information extracted from orthoimages and observation-related data in a similarity analysis unit, and FIG. 8 is a geometric model between reference data and observation-related data in a network model construction unit It is a diagram showing a process of deriving correction data from the most accurate value calculation unit while generating. FIG. 9 is a diagram illustrating processing for calculating 3D geometric data by constructing an intersection geometry from a multi-view or stereo view according to a movement trajectory of a mobile platform, and FIG. 11 is observation-related data acquired by a 3D survey sensor A is a diagram representing geometrically corrected correction data based on a normal image.

First, in FIG. 3, the sensor unit mounted on the mobile platform acquires sensor data from at least one sensor 102 to 106, and the sensor data processing unit 110 processes sensor data to acquire a number of observation related data. Then, the orthogonal image searcher 108 searches for a plurality of orthogonal data of the orthogonal image 16 based on the observation-related data (S305).

Observation-related data is calculated based on observation data output directly from each sensor (102~106), external geometry defining the geometric relationship of the sensor according to the position and attitude of the mobile platform, and geometric parameters uniquely defined for each sensor. It may include at least one of the internal geometry. The external geometry and the internal geometry can constitute geometric information.

In addition, the observation-related data includes the above-described observation data, geometric information, time information for the acquisition time of each sensor, unique data related to device-specific attributes for each sensor 102-106, point group based on the three-dimensional survey sensor 104 Based on at least one of data-related information, geocoding image data-related information based on the image sensor 102, point group data-related information, and geocoding image data-related information, object information, point group estimated by attribute of the extracted object Based on at least one of data-related information and geo-coded image data-related information, at least one of map-related information generated to create a map may be included.

1 and 5, as shown in FIGS. 1 and 5, the orthogonal image search unit 108 retrieves and extracts a plurality of orthogonal data as orthogonal image related data for a region of interest corresponding to the movement trajectory 18 of the mobile platform from the orthogonal image 16 do.

Ortho data has geometric shape information estimated to be feature geometry in the region of interest, and the geometric shape information may be extracted based on a specific geometric shape of the object extracted in the region of interest. 5, the object may be randomly selected as an object capable of easily extracting feature geometries such as points, lines, faces, and columns on a map, such as a road sign, an object of unusual shape, and a building boundary line.

As in this embodiment, by setting the region of interest that matches the movement trajectory 18 of the mobile platform to search for and extract ortho data, contrast with observation-related data processed from sensor data obtained from the sensors 102 to 106 Not only can the time required to set the reference data and analyze the similarity be shortened, but also the accuracy in the process of estimating the correction data can be improved.

Next, the similarity analysis unit 112 sets at least one reference data that is equally estimated from both data by similarity analysis between the searched orthogonal data and observation-related data (S310).

The similarity analysis unit 112 may perform a predetermined preprocessing on at least one of acquisition observation related data and orthogonal data from at least one sensor 102 to 106 before the similarity analysis.

The above-described pre-processing is at least one of pre-processing of image luminance adjustment, pre-processing of image resolution adjustment, pre-processing of image sharpness adjustment, pre-processing of image geometric transformation, pre-transformation of image coordinate system, pre-extraction of object boundary information from observation and ortho-image related data. Can be applied. This pre-processing improves the accuracy of similarity analysis by minimizing the amount of deformation such as geometric information, coordinate system, and resolution due to differences in the observed data characteristics such as the unique data of the sensors 102 to 106, the environment during observation, and scale/rotation/displacement values. To be carried out. For example, an orthogonal image of orthogonal data is acquired at noon and clear weather, and when observation-related data is acquired at rainy weather at sunset, the resolution of both data may be different and errors in similarity analysis may be caused. To solve this, pre-processing of adjusting the resolution of both data to a close value may be performed. The image geometric transformation pre-processing may be image scale transformation or image rotation change.

In the similarity analysis performed after the pre-processing, when obtaining observation-related data from the image sensor 102, the similarity analysis unit 112 changes color based on pixel coordinates among geocoding image data-related information belonging to the observation-related data By analyzing, it is possible to perform similarity analysis to extract geometric shape information and set reference data. When obtaining observation-related data from the 3D survey sensor 104, the similarity analysis unit 112 at least one of the geometric shape change and the laser intensity change of the point cloud data belonging to the observation-related data as shown in FIG. By analyzing, it is possible to perform similarity analysis to extract geometric shape information and set reference data. According to FIG. 6, a process of setting the reference data by performing similarity analysis when the feature geometry is geometric, column, and point information of the observation data and orthogonal data of the Lidar sensor is shown. 7 shows a process of setting reference data by performing similarity analysis using points, boundary information, etc., as a feature geometry of observation-related data and ortho data obtained from the image sensor 102.

According to this, it is possible to improve the matching precision by analyzing the similarity between the data by using geometric shape information such as a specific geometric shape or a type of object as object information of the reference data and observation-related data.

Next, the network model construction unit 114 generates a geometric model between reference data and observation-related data based on the reference data and the reference coordinate system 20 of the orthoimage (S315).

Specifically, the network model construction unit 114 generates a geometric model between reference data and the observation-related data based on geometric information composed of external geometry and internal geometry. When the correction data represents an unstable value in the most accurate value calculating unit 116, a geometric model is generated through an optimization process that considers internal geometry included in reference data and observation related data.

The network model constructing unit 114 may calculate an objective function that minimizes the distance from observation-related data through a geometric model to which a plurality of parameters are applied.

In FIG. 8, the network model constructing unit 134 configures a geometric model in terms of reference data and a reference coordinate system by geometric information including internal geometry and external geometry.

Specifically, the detailed process of calculating the internal geometry for each sensor 102 to 106 is exemplified below.

In the geometric model of FIG. 8, the internal geometric model of the 3D measurement sensor 104 may be defined as [Equation 1].

[Equation 1]

Also, in the geometric model of FIG. 8, the internal geometric model of the image sensor 102 may be defined as [Equation 2].

[Equation 2]

In addition, the external geometric model of the three-dimensional survey sensor 104 for data fusion by applying the internal geometric model of the defined three-dimensional survey sensor 104 and the internal geometric model of the defined image sensor 102 is shown. In the geometric model of 8, it is defined as in [Equation 3], and the external geometric model of the image sensor 102 can be defined as in [Equation 4].

[Equation 3]

[Equation 4]

As another embodiment, if the correction data in the most accurate value calculation unit 116 shows a stable value, the network model construction unit 114 generates a geometric model between reference data and observation-related data based on external geometry among the geometric information. Can. Considering that the positions and postures of the mobile platform on which the sensors 102 to 106 that acquire observation-related data are mounted and the sensors that acquire the orthoimage are different from each other, external geometry may be preferentially used when generating the geometric model.

The network model constructing unit 114 may apply the reference data without vertical geometric data when constructing the geometric model, but the orthonormal data is two-dimensional data, and observation-related data obtained from the three-dimensional survey sensor 104 is 3 In the case of dimensional data, in order to increase the consistency of the geometric model, a supplementary process of adding vertical geometric data to the reference data applied to the geometric model may be further included.

In this case, the complementary processing is a process of setting vertical geometric data to a predetermined value, a process of inverting three-dimensional coordinates with a large number of reference data, a vertical geometry data in a raster form, a high-resolution model or a vector form, It may be at least one of the process of adding 3D to the reference data and the process of calculating the 3D geometric data by the intersection geometry 30.

The processing by the intersection geometry is performed according to the movement trajectory 18 of the mobile platform, as shown in FIG. 9, where the sensors 102-106 are acquired at multiple points in multi-view or stereo view. Vertical geometric data is calculated by constructing an intersection geometry 30 with respect to the feature geometry 22 of reference data photographed in regions overlapping each other in the observation range 28 for each point from, This is a process of adding data to the reference data 24 of the orthogonal image 24. In the case where the geometric stability is not secured by the complementary processing by the intersection geometry, the other processing described above may be additionally applied.

Next, the maximum value calculating unit 116 calculates a maximum value that minimizes the amount of separation between observation-related data and reference data in the geometric model, and estimates correction data from the observation-related data according to the maximum value (S320).

The most accurate value calculating unit 116 uses a predetermined analysis model formula, and some parameters among a plurality of parameters may be applied to the model formula. The maximum value may be at least one of external geometric and internal geometric error information. In addition, the maximum value may be calculated as at least one of error information of geometric shape information generated by at least one of the above-described information, point group data related information, geocoding image data related information, object information, and map related information.

The maximum value may be calculated using numerical analysis that statistically minimizes the separation between reference data such as object information, coordinate data, geometric information, and time information and related information belonging to the collected data, and the maximum value is a scalar value, It can be expressed in the form of a matrix or model.

For example, in order to apply the least squares method as a process of finding the maximum value, the data geometric model must be Gauss-Markov modeled, which can be defined by the following equation.

[Equation 5]

는 관측값,

는 기본 기하식,

는 기존 기하 모델 파라미터,

는 기본 기하식과 파라매터 간 관계에서 유도된 자코비안(Jacobian) 행렬,

는 기하 모델 파라미터 보정량,

는 부정 오차이다.here,

Is the observation,

Is the basic geometry,

Is the existing geometric model parameter,

Is a Jacobian matrix derived from the relationship between the basic geometrical and parametric,

Is the geometric model parameter correction amount,

Is the negative error.

The correction amount of the geometric model parameter can be inversely calculated using Gauss-Markov, which can be defined by the following equation.

[Equation 6]

는 역계산된 기하 모델 파라미터 보정량으로서 최확값이고,

는 관측값의 정확도에 따라 설정되는 관측값의 경중률이다.here,

Is the inverse computed geometric model parameter correction and is the maximum value,

Is the weight ratio of the observations set according to the accuracy of the observations.

이 수집 데이터에 대해 병합함으로써, 도 7에서와 같이 보정 데이터(RL_1)가 생성될 수 있다. Maximum

By merging the collected data, correction data RL_1 can be generated as in FIG. 7.

The related data of the correction data may include variance and covariance values thereof together with the above-described values.

By estimating the correction data with the calculated maximum value, geometric correction of observation-related data using the reference data is performed, and such time-limit correction is the result of applying the reference data to the geometric model and the geometry of observation-related data corresponding to the reference data This is the process of analyzing the error amount and calculating the minimum or optimized parameter as the maximum value.

An image in which the correction data is estimated according to the system 100 of this embodiment and the geometric correction related to the sensor unit mounted on the mobile platform is made is illustrated through FIG. 11. In this case, the 3D observation-related data obtained from the Lidar sensor mounted on the mobile platform is geometrically corrected based on the orthoimage, and the Lidar data is displayed in blue, and the movement trajectory of the mobile platform is displayed in red.

According to the present embodiment, by utilizing the feature geometry in the orthoimage to analyze the similarity between the observation-related data and the orthoimage of the sensor mounted on the mobile platform, thereby generating a geometric model between both data based on the reference data , It is possible to secure accuracy and stability of 3D geometric information correction of a mobile platform using an orthogonal image composed of 2D terrain information.

According to the present embodiment, the orthoimage is easy to secure and utilize the reference data in a one-to-one relationship with the coordinates of the ground, and can facilitate management of the reference data when performing geometric correction of the mobile platform. Based on this, it is possible to generate highly accurate correction data in correction of observation related data.

Hereinafter, with reference to FIGS. 2 and 10, the process of processing in the case where the expected error according to the maximum value exceeds the allowable error in the system according to an embodiment of the present invention will be described.

10 is a flowchart illustrating a processing process when an expected error according to a maximum value exceeds an allowable error in a geometric correction system of a mobile platform equipped with a sensor according to an embodiment of the present invention.

First, the predicted error probability analysis unit 118 analyzes the predicted error between the correction data and the observation-related data based on the minimum separation amount used when calculating the maximum value (S1005), and whether the predicted error exceeds the allowable error It is determined (S1010).

In case of exceeding, at least one of the following processes (at least) of the corresponding unit (module), that is, an orthogonal image search unit 108, a similarity analysis unit 112, a network model construction unit 114, and a maximum value calculation unit 116 It is controlled by instructing any one, and restarting the correction process from the relevant part (module) related to the newly set constraint, so as to regenerate the maximum value and correction data (S1015).

The predicted error probability analysis unit 118 searches for orthoimages to perform a process of searching for and extracting orthonormal data by adjusting a region of interest that matches the moving trajectory 18 of the mobile platform when the predicted error exceeds the allowable error. The unit 108 can be controlled. For example, if the initial region of interest is too wide or the feature geometry is selected less than the city center, such as a mountainous region, the region of interest may be set to be narrow, but the feature region may be set to a relatively large region.

The predicted error probability analysis unit 118 changes at least one of the pre-processing performed by the similarity analysis unit 112, or uses the geometric shape information estimated as another object or other feature geometry in orthogonal data and observation-related data as a new criterion. The similarity analysis unit 112 may be controlled to perform a process of sampling and extracting data. For example, image luminance adjustment pre-processing, image resolution adjustment pre-processing, image sharpness adjustment pre-processing, image geometric transformation, image coordinate system transformation, and pre-processing for extracting object boundary information from observation and ortho-image related data may be performed.

The predicted error probability analysis unit 118 may control the network model construction unit 114 to perform a process of determining and removing observation-related data representing a predetermined distance from the objective function calculated through the geometric model as abnormal data. . For example, if there is observation-related data with a high degree of separation between the objective function and the observation-related data (L_1) of the first object and the observation-related data (L_2) of the second object, the data is determined as abnormal data and removed. In some cases, reference data related to this may be excluded.

The predicted error probability analysis unit 118 may control the network model construction unit 114 to perform a process of changing at least some parameters applied to the geometric model.

The predicted error probability analysis unit 118 is configured to perform at least one of the process of changing at least some parameters used in the analysis model formula of the most applicable value applied to the most accurate value calculation unit 116 or changing the analysis model formula. It is also possible to control the value calculation unit 116.

If the expected error does not exceed the allowable error, the correction data estimated in step S320 is finally adopted.

According to the present embodiment, by correcting the correction process of FIG. 3 to generate more reliable correction data, the accuracy and stability of the geometric correction can be further improved.

The components constituting the system 100 shown in FIG. 2 or steps according to the embodiments shown in FIGS. 3 and 10 may be recorded in a computer-readable recording medium in the form of a program realizing the function. Here, a computer-readable recording medium refers to a recording medium that can store information such as data or programs by electrical, magnetic, optical, mechanical, or chemical action, and can be read by a computer. Examples of such recording media that can be separated from a computer include, for example, portable storage, flexible discs, magneto-optical discs, CD-ROMs, CD-R/Ws, DVDs, DATs, and memory cards. In addition, there are solid state disks (SSDs), hard disks, ROMs, etc., as recording media fixed to mobile devices and computers.

In addition, in the above, even if all components constituting the embodiments of the present invention have been described as being combined and operated, the present invention is not necessarily limited to these embodiments. That is, within the scope of the object of the present invention, all of the constituent elements may be selectively combined and operated. In addition, although all of the components may be implemented by one independent hardware, a part or all of the components are selectively combined to perform a program module that performs a part or all of functions combined in one or a plurality of hardware. It can also be implemented as a computer program.

Although the present invention has been described in detail through exemplary embodiments above, those skilled in the art to which the present invention pertains are capable of various modifications to the above-described embodiments without departing from the scope of the present invention. Will understand. Therefore, the scope of rights of the present invention should not be limited to the described embodiments, but should be determined by any modified or modified form derived from the claims and equivalent concepts as well as the claims described below.

100: system 102: image sensor
104: three-dimensional measurement sensor 106: navigation sensor
108: orthographic image search unit 110: sensor data processing unit
112: similarity analysis unit 114: network model configuration unit
116: most accurate value calculation unit 118: expected error probability analysis unit

Claims (10)

In the geometric correction system of the mobile platform with a sensor based on orthoimage,
A sensor unit including at least one of an image sensor, a navigation sensor, and a 3D measurement sensor for obtaining 3D geographic data;
Similarity analysis is performed between observation-related data obtained from at least one sensor mounted on the mobile platform and orthoimage-related data having two-dimensional terrain information, and reference data estimated equally from the observation-related data and the orthoimage-related data. Similarity analysis unit for setting;
A network model construction unit generating a geometric model between the reference data and the observation-related data based on the reference data and a reference coordinate system of the orthoimage;
A most probable value that calculates a most probable value that minimizes the amount of separation between the observation-related data and the reference data in the geometric model and generates correction data estimated from the observation-related data according to the maximum value; And
And a predicted error probability analysis unit that analyzes a predicted error between the correction data and the observation-related data based on the minimum distance used when calculating the maximum value, and determines whether the predicted error exceeds the allowable error,
When the expected error probability analysis unit exceeds the expected error, the allowable error,
A process of retrieving and extracting the orthographic image-related data by adjusting a region of interest that matches the movement trajectory of the mobile platform,
When the similarity analysis unit performs pre-processing to minimize deformation due to a difference in observed geometric characteristics between the observation-related data and the ortho data before the similarity analysis, the conditions of the pre-processing are changed, or the ortho-image data and the observation-related A process of sampling and extracting geometric shape information estimated from another object or another feature geometry from data as new reference data,
A process of removing observation-related data determined as abnormal data from the geometric model generated by the network model construction unit;
A process of changing at least some parameters applied to the geometric model, and
The similarity analysis unit and the network model construction unit to perform at least one of the process of changing at least some parameters used in the analysis model formula of the most applicable value applied to the most accurate value calculation unit or changing the analysis model expression. , Control the most accurate value calculation unit,
A geometric correction system of a mobile platform equipped with a sensor that controls to generate the maximum value and correction data again according to constraints after performing the processing. According to claim 1,
The orthogonal image is image information reflecting an orthogonal projection model in which distortion caused by the projection geometric model of the camera is removed,
The observation-related data is based on observation data directly output from each sensor, external geometry defining a sensor's geometric relationship according to the position and posture of the mobile platform, time information on the acquisition time of each sensor, and the three-dimensional survey sensor. Point group data-related information, geocoding image data-related information based on the image sensor, object information for which an attribute of the extracted object is estimated based on at least one of the point group data-related information and the geocoding image data-related information, Based on at least one of the point group data related information and geocoding image data related information, a geometric correction system of a mobile platform equipped with a sensor including at least one of map related information generated to create a map. According to claim 2,
The observation-related data further includes an internal geometry calculated based on geometric parameters uniquely defined for each sensor,
The network model construction unit is a geometric correction system of a mobile platform equipped with a sensor that generates a geometric model between the reference data and the observation-related data based on the geometric information composed of the external geometry and the internal geometry. The method of claim 3,
The maximum value is the error information of the geometry information generated by at least one of the external geometry, the error information of the internal geometry, the point group data related information, the geocoding image data related information, the object information, and the map related information. At least one selected from, and the geometric shape information is information related to the feature geometry estimated to be the same from the reference data and the observation related data. A geometric correction system for a mobile platform equipped with a sensor. According to claim 1,
Further comprising an orthogonal image search unit for retrieving and extracting orthogonal data for a region of interest corresponding to the movement trajectory of the mobile platform from the orthogonal image,
The orthoimage-related data includes orthogonal data having geometric shape information estimated to be feature geometry in the region of interest, and the geometric shape information is a sensor extracted based on a specific geometric shape of an object extracted in the region of interest Geometry correction system on the mobile platform. The method of claim 5,
The similarity analysis is performed to set the reference data that is estimated to be equal between the orthonormal data and the observation-related data, and when the similarity analysis unit acquires the observation-related data from the image sensor, geo belonging to the observation-related data When analyzing the color change based on pixel coordinates among the information related to coded image data, extracting the geometric shape information and performing similarity analysis to set the reference data, and obtaining the observation-related data from the 3D survey sensor , A mobile platform equipped with a sensor that analyzes at least one of the geometric shape change and laser intensity change of the point cloud data belonging to the observation-related data to extract the geometric shape information and perform similarity analysis to set the reference data Geometric correction system. According to claim 1,
The similarity analysis unit prior to the execution of the similarity analysis, at least one of the observation-related data and the orthoimage-related data, image luminance adjustment pre-processing, image resolution adjustment pre-processing, image sharpness adjustment pre-processing, image geometry transformation pre-processing, image coordinate system conversion A geometric correction system of a mobile platform equipped with a sensor that performs at least one pre-processing selected from pre-processing, pre-processing for extracting boundary information of an object from the observation and ortho-image related data. delete According to claim 1,
The network model construction unit further includes a supplementary process of adding vertical geometric data to the reference data applied to the geometric model, or a sensor mounted to apply the reference data without the vertical geometric data when constructing the geometric model Correction system of the old mobile platform. The method of claim 9,
The complementary processing includes processing of setting the vertical geometric data to a predetermined value, processing of inverting three-dimensional coordinates into a plurality of reference data, and raster-shaped digital elevation model or vector-shaped numerical geometry data. Depending on the processing added to the reference data and the movement trajectory of the mobile platform, the sensors overlap each other in a multi-view or stereo-view observation range for each of the multiple points obtained from multiple points. A geometric correction system for a mobile platform equipped with a sensor, which is performed by at least one of a process of calculating the vertical geometric data by configuring an intersection geometry of a region to be formed.

Images