JP2016001182A - Three-dimensional scanning beam system and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a three-dimensional scanning beam system and method, which enable economical and efficient three-dimensional scanning.SOLUTION: A system 100 includes a ranging apparatus 105, and a reactive linkage mechanism 110 having a first end 115 and a second end 120. The first end 115 is connected to the ranging apparatus 105 and the second end 120 is connected to an object 125 that moves the system 100 through an environment. In use, acceleration of the object 125 with respect to the environment is converted by the reactive linkage mechanism 110 to motion of the ranging apparatus 105 with respect to the object 125, which increases the field of view of the ranging apparatus 105 with respect to the environment.

Description

  The present invention relates generally to scanning beam systems. Although not limiting, the present invention is particularly related to systems and methods for three-dimensional (3D) scanning using a reactive coupling mechanism.

  A vehicle, person, or robot operating in an environment where a priori maps are not available to function and adapt to its work environment or to provide an accurate model to the end user Often, you have to build a map. Map creation and self-position estimation, including SLAM (simultaneous localization and mapping), is a fundamental problem in mobile robotics, moving objects that do not have a known map or an accurate measure of their pose Includes a scenario where both must be estimated and maintained.

  The posture of the object refers to the position and orientation of the object in space. Knowing exactly the pose of a moving object is often important for mobile tasks such as robot navigation, map creation, and planning. In particular, it is well known that in some environments conventional navigation systems cannot operate or are impractical. For example, in environments such as indoors, underground mines, or deep seas, satellite-based global positioning system (GPS) technology and cellular-based positioning technology often do not work. If a detailed map of such an environment is not available, a vehicle, person, or robot operating in that environment must rely on an alternative navigation system, such as a SLAM system, An inertial measurement unit (IMU) or other sensor may be included to determine position and orientation.

  The 3D scanning beam system can provide data for map creation and / or self-position estimation applications. In general, such systems include an active motorized linkage that moves the sensor in multiple degrees of freedom to effectively increase the field of view of the sensor to obtain 3D scanning. For example, some systems may include a laser ranging device that outputs a one-dimensional beam. When the distance measuring device is mounted on an electric multi-degree-of-freedom coupling mechanism or an electric gimbal mechanism, the laser can be repeatedly swept in a three-dimensional space such as a space in front of the vehicle. The data collected from the device can then be used to define a 3D “point cloud”. A 3D “point cloud” effectively defines an image or map of the environment. Other ranging devices can comprise a two-dimensional laser scanner, which can be rotated about the central axis or moved in the third dimension when fixed to the vehicle. A 3D point cloud is defined. Furthermore, a 3D flash light detection and ranging (LIDAR, light detection and ranging) device can incorporate a plurality of detectors in parallel to generate 3D ranging data.

  In many cases, the 3D scanning beam system is fixedly mounted on the vehicle or robot and continuously scans a specific area relative to the orientation of the vehicle or robot. The collected data can then be used for mapping and / or self-location estimation applications, or for more basic object detection and scene recognition purposes. However, in some situations, the system can be errored due to the vibration or up and down vibrations of the vehicle or robot to which the fixedly mounted scanning beam system is attached. Thus, complex vibration isolation and stabilization systems have been devised to effectively decouple the 3D scanning beam system from unwanted vibrations or motion.

  In some applications, the hardware required to effectively scan a three-dimensional region by increasing the field of view of the scanning sensor can be extremely complex and expensive. Accordingly, there is a need for improved 3D scanning beam systems and methods.

Michael Bosse and Robert Zlot, “Continuous 3D Scan-Matching with a 2D Spinning Laser,” Proceedings of the IEEE International 43

  The purpose of some embodiments of the present invention is to provide consumers with improvements and advantages over the prior art described above, and / or to overcome and mitigate one or more of the above-mentioned drawbacks of the prior art, And / or to provide useful commercial options.

  Accordingly, in one form, the present invention comprises a distance measuring device and a reaction coupling mechanism having a first end and a second end, wherein the first end is connected to the distance measuring device. The second end is connected to an object that moves the system in the environment, and in use, the acceleration of the object relative to the environment is converted into a movement of the distance measuring device relative to the object by the reactive coupling mechanism in use. Thus, the field of view of the distance measuring device with respect to the environment is increased.

  Optionally, the system further comprises an orientation sensor attached to the ranging device.

  Optionally, the system further comprises a computer system connected to the ranging device, the computer comprising a computer readable medium storing a program code component defining a registration algorithm, wherein the registration algorithm Process data from the ranging device to define a 3D point cloud of the environment.

  Optionally, the ranging device comprises a laser.

  Optionally, the laser comprises a two-dimensional scanning beam laser.

  Optionally, the 3D point cloud is defined in real time.

  Optionally, the ranging device transmits and receives at least one of the following types of signals: optical signal, acoustic signal, ultrasonic signal, radio frequency signal, γ radiation signal, microwave signal.

  Optionally, the reaction coupling mechanism comprises a spring.

  Optionally, the orientation sensor comprises an inertial measurement unit (IMU).

  Optionally, the connection between the first end of the reactive coupling mechanism and the ranging device is a fluid coupling.

  According to another aspect, the present invention is a method for obtaining a three-dimensional scan of an environment, the method comprising connecting a ranging device to a first end of a reactive coupling mechanism, said reactive coupling mechanism Connecting the second end of the object to the object, moving the object in the environment, wherein the acceleration of the object is converted into a movement of the ranging device relative to the object by the reaction coupling mechanism. , Increasing the field of view of the ranging device relative to the environment, and obtaining data defining the three-dimensional scan of the environment by the ranging device.

  Optionally, the method further comprises connecting a computer system comprising a computer readable medium having stored program code components defining a registration algorithm to the ranging device and using the computer system. Processing the data, the registration algorithm comprising defining a 3D point cloud of the environment.

  Still other features and advantages of the present invention will become apparent from the following detailed description.

  In order to assist the understanding of the present invention and to enable those skilled in the art to practice the invention, preferred embodiments of the invention will now be described by way of example only with reference to the accompanying drawings.

1 is a schematic diagram illustrating aspects of a 3D scanning beam system according to an embodiment of the present invention. FIG. FIG. 2 is a schematic diagram showing a side view of the system of FIG. FIG. 2 is a schematic diagram showing a side view of the system of FIG. 1 connected to the top of a construction safety helmet. FIG. 6 illustrates the translation of the scanner relative to a surface in the environment, according to some embodiments of the invention. 2 is an overall flow diagram illustrating a method for obtaining a three-dimensional scan of an environment according to one embodiment of the invention. FIG. 2 is a block diagram illustrating components of a system including a computer system for defining a 3D point cloud according to some embodiments of the present invention. FIG. 6 is a schematic diagram illustrating aspects of a 3D scanning beam system according to an alternative embodiment of the present invention.

  The present invention relates to 3D scanning beam systems and methods. By attaching this system to objects such as vehicles, robots, people or animals, 3D scanning of the environment can be obtained economically and reliably. The data from the 3D scan can then be used for various purposes such as object detection, proximity detection, map creation, self-position estimation, or collision avoidance.

  In this patent specification, adjectives such as first and second, left and right, up and down are used only to define one element or method step distinct from another element or method step; The specific relative position or order described by these adjectives is not necessarily required. Terms such as “comprising” or “including” are not intended to define an exclusive set of elements or method steps. More specifically, such terms only define a minimal set of elements or method steps included in a particular embodiment of the invention.

  Referring to FIG. 1, a schematic side view of a 3D scanning beam system 100 according to an embodiment of the present invention is shown. The system 100 comprises a ranging device in the form of a 2D laser beam scanner 105 and a reactive coupling mechanism in the form of a coil spring 110. The first end 115 of the coil spring 110 is connected to the 2D laser beam scanner 105. The second end 120 of the coil spring 110 is connected to the object 125. The translational acceleration of object 125 (indicated by arrow 130) is then converted into a rotational motion of 2D laser beam scanner 105 (indicated by arrow 135) by a reactive coupling mechanism.

  Thus, the 2D laser beam scanner 105 can significantly increase its field of view and scan the environment in front of the object 125 in 3D. An independently powered gimbal mechanism or other independent excitation source is not required to produce the rotational motion of the 2D laser beam scanner 105. Thus, the system 100 can be economical and effective when cost or space constraints impose limitations on the hardware required to scan the environment using a ranging device.

  According to some embodiments of the present invention, an optional orientation sensor, such as an inertial measurement unit (IMU) 145, camera, magnetometer, or other device, can be attached to the ranging device. Such an orientation sensor can assist in processing data from a ranging device in applications where data defining the location or orientation of the ranging device is required.

  Referring to FIG. 2, a schematic diagram of a side view of the system 100 connected to the front end of an automobile 200 is shown. Therefore, the object 125 is replaced with the automobile 200. The translational acceleration of the automobile 200 is converted into the rotational motion of the 2D laser beam scanner 105 so that the area in front of the automobile 200 can be scanned in 3D. Other naturally occurring accelerations of the automobile 200 in normal use, such as up and down (ie up and down) acceleration and turning acceleration, can help excite the rotational motion of the 2D laser beam scanner 105. As will be appreciated by those skilled in the art, the force and torque applied to the object 125 induces a movement of the scanner 105 relative to the object 125 (primarily rotational, but also with small translational movement). As a result, the movement of the scanner 105 relative to the environment being scanned is amplified.

  2D laser beam scanners such as scanner 105 are either widely available commercially or can be easily manufactured using commercially available components. For example, the scanner 105 may include a light detection and ranging (LIDAR) laser measurement sensor (LMS, laser measurement sensor) model 291 manufactured by SICK AG. Alternatively, various methods, including those using various types of radiation or signals, such as optical signals, acoustic signals, ultrasound signals, radio frequency (RF) signals, gamma radiation signals, or microwave signals, for example Other types of ranging devices can also be used. Such radiation or signals may be used in various ranging techniques including, for example, time of flight (TOF), frequency modulated continuous wave (FMCW), triangulation, and phase shift ranging techniques. Can do. Further, the distance measuring device according to the present invention does not need to be scanned in 2D, and can be a simple one-dimensional beam distance measuring device, for example. Such rotation of the one-dimensional beam ranging device can also provide effective 3D scanning of the environment.

  According to some embodiments of the present invention, the reactive coupling mechanism may also include various types of commercially available off the shelf (COTS) products or custom made products. For example, the coil spring 110 may include various types of coil springs made of various materials including polymers and metal alloys. Further, the reaction coupling mechanism may include other elements such as a plurality of springs connected to the distance measuring device and a spring-like material. As will be appreciated by those skilled in the art, in view of this disclosure, various other types of elastic connection mechanisms are used, including, for example, simple elastic rods, inflated balloons, sponges, rubbers, etc. A reactive coupling mechanism that converts the acceleration of the object into the movement of the distance measuring device can be provided.

  Further, those skilled in the art will appreciate that the reactive coupling mechanism according to the present invention can also act as a passive mechanical filter that isolates the scanner 105 from the high frequency motion of the object 125. A reactive coupling mechanism, such as coil spring 110, can be adjusted to both resonate at a specific frequency to meet application requirements and to suppress high frequency motion that may be problematic. it can. For example, some high frequency motion can complicate processing by the registration algorithm or be detrimental to the scanner 105.

  Further, the reactive coupling mechanism can be adjusted to increase the impulse response of the scanner 105 based on the time constant associated with the hardware so that the scanner 105 moves over a period of time after the object 125 stops moving. Effectively making it possible to continue. Thus, the reactive coupling mechanism can accumulate energy (such as mechanical energy or gravitational energy) and slowly release it over time.

  Still other embodiments of the reactive coupling mechanism may include a damper or other physical limiter of the range of motion available to the mechanism. This can be useful, for example, to limit the field of view of the scanner 105 to a range of interest for a particular application.

  Referring to FIG. 3, a schematic side view of the system 100 connected to the top of the construction safety helmet 300 is shown. According to this embodiment, the movement of the head of the worker wearing the helmet 300 due to walking, whirling movement, or other natural movement is used to excite the rotational movement of the 2D laser beam scanner 105. Can be sufficient. Thus, the system 100 can be used to effectively scan a work site immediately in front of an operator, such as indoors, underground mines, or various other environments.

  The system 100 shown in FIG. 3 can be designed to use very little power because no motor or other motorized mechanism is needed to excite the scanning or rotational movement of the ranging device. Thus, the distance measuring device need only be powered by a relatively low power battery that is required, for example, to supply power to the solid state circuit of the distance measuring device.

  Some embodiments of the present invention may be used to provide scan data to a computer system. The computer system can then process this data using a registration algorithm to define a 3D point cloud. As known to those skilled in the art, such a 3D point cloud can be used in various map creation and / or self-position estimation applications or imaging applications.

  Thus, a system such as system 100 can be used in various applications. For example, in the field of map creation, the system 100 can be mounted on a mobile platform such as a vehicle, a wheelbarrow, a human, or an animal and can be used to obtain a 3D map of the environment. In robotics applications, the system 100 can be used for terrain mapping or general scene recognition. As shown in FIG. 3, the various lightweight embodiments of the present invention can be used as wearable or handheld devices for navigation purposes. For example, the system 100 may be used by first-time personnel, such as police, firefighters, or military personnel, within a building that has been mapped. The locally acquired map created from the data from the system 100 can then be compared in real time with the existing global map to locate and assist the initial personnel. Similar self-position estimation and map creation functions can be achieved using robots or other vehicles in remote environments such as underground mines and underwater environments.

  An example of a registration algorithm that can be used with the present invention is provided in a paper that is incorporated herein by reference in its entirety (see, for example, Non-Patent Document 1). A summary of how such a registration algorithm works is described below.

  For example, considering the system 100 connected to the helmet 300 shown in FIG. 3, as the person wearing the helmet 300 moves, motion is excited through the spring 110. Thus, the scanner 105 rotates over additional degrees of freedom and a wider field of view, thereby taking measurements outside the normal 2D scan plane to provide 3D measurements of the surrounding environment. However, the exact nature of the scanner motion may not be known at first. Although data from the IMU 145 fixedly attached to the laser scanner 105 can be used to estimate the rotational motion of the scanner 105, the translational motion may not be known at all. A registration algorithm can recover translational motion, correct the rotational motion estimates, and perform additional timing synchronization.

Referring to FIG. 4, the translational movement of the scanner 105 relative to the surfaces z ₁ and z ₂ in the environment is shown according to some embodiments of the present invention. As shown, consider that scanner 105 is in the first position at time t ₁ and is in the second position at time t ₂ . Before the data obtained by the system 100 is processed using a registration algorithm, the attitude (ie position and orientation) of the scanner 105 at times t ₁ and t ₂ is not known accurately (some estimate of position). For example, based on data from the IMU 145). As illustrated, both surfaces z ₁ and z ₂ represent surfaces in the environment that are measured by beam 140 from scanner 105 at both first and second locations. The first step in the registration algorithm is to perform a data association that identifies that the scanner 105 is actually looking at a common feature at both locations that identify the surfaces z ₁ and z ₂ . A constraint system H is then formed that links the feature match to the correction x applied to the attitude of the scanner 105. As will be appreciated by those skilled in the art, only a subset of postures and matches is shown in FIG. In contrast, actual embodiments will generally consider a longer trajectory of the scanner 105 and a complete set of range measurements.

  The registration algorithm can use as input the previous trajectory (eg, including only orientation estimates taken from IMU 145) and a set of range scans. Using the previous trajectory, the scan points (ie range measurements) can be projected onto an unregistered (ie generally less aligned) 3D point cloud. The trajectory can be defined as a continuous time history of the origin position and orientation of the scanner 105 (in six degrees of freedom (DOF), ie, three positions DOF and three rotational DOFs). The posture is defined by 6 DOFs for the 3D rigid body. Thus, the trajectory can be defined as a function T (τ) that specifies the orientation of the six DOFs of the scanner 105 relative to the ground coordinate frame for each time τ over the duration of the sweep of the scanner 105.

The environment workspace is then discretized into a 3D grid of cells or “voxels” (volume elements), and statistics are calculated for each voxel based on the set of scan points contained within its boundaries. To reflect the non-uniform sampling of the environment and boundary effects, the 3D grid is calculated at multiple resolutions and offsets. The statistics calculated for a given voxel define the ellipsoids that fit these points using the first and second moments μ, S that describe the parameters for the ellipsoid that fits the points p _i .

  Ellipsoids can be summarized by their centroids (ie, average positions) and a set of coordinate axes and sizes that define the shape (which can be computed by the eigenvalue-eigenvector decomposition of the second moment matrix S). The shape can additionally be summarized by shape parameters that describe how flat, cylindrical, or sphere the distribution is based on the calculated eigenvalues. The plane can be further associated with its normal (the eigenvector corresponding to the smallest eigenvalue) and the cylinder can be associated with its axis (the eigenvector corresponding to the largest eigenvalue).

は、０〜１の範囲の、ポイントの円柱らしさである。同様に、量

は、０〜１の範囲の、ポイントの平面らしさである。 Given a sorted eigenvalue λ ₁ ≦ λ ₂ ≦ λ _{3 of the second} moment matrix S, the quantity

Is the cylindrical character of the point in the range of 0-1. Similarly, quantity

Is the flatness of the point in the range of 0-1.

  The sensor trajectory is then estimated using an iterative optimization algorithm and the point cloud is registered to create an accurate 3D map. At each iteration, an input trajectory estimate is provided and further corrected to produce an improved output trajectory estimate. This estimate can then be used as an input for the next iteration. The first iteration uses the previous estimate generated from the rotational motion estimated by IMU 145. The registration algorithm can be applied to the trajectory segments at a fixed interval (eg, once per second) to register the most recent data as data is obtained from the ranging device and orientation sensor. By accumulating registered trajectory segments in this way (and maintaining continuity constraints between segments), a complete trajectory can be obtained little by little.

  Registration is performed as a minimization problem with a linear system of constraints. To form such a system, the trajectory must be discretized by sampling the trajectory at regular intervals. Later, interpolation can be used to form a continuous trajectory given a discrete solution. Optimization provides a correction solution for the trajectory, resulting in a smooth trajectory that best represents the range data.

  The algorithm works in turn on fixed-size overlapping time segments of the trajectory. For example, the first second of data is processed, then the next second of data (partially overlapping with the first second of data).

  At each iteration of the registration algorithm, a set of constraints is generated. Each constraint is one of three types. The first type includes a smoothness constraint, which is how the scanner 105 (and hence the object to which the scanner 105 is connected) can move based on physical constraints such as smooth motion assumptions, ie progressive acceleration limits. Incorporate a model for whether there is sex. The second type of constraint includes an initial condition constraint, which ensures that the current trajectory segment is continuous with the previous segment (ie, the start of the current segment coincides with the end of the previous segment). Should be constrained to).

  The third type of constraints include match constraints, which take into account laser measurements of the external environment. In this case, related voxel pairs are identified. These voxels are presumed to be matches originating from the same target in the environment, measured at different times. Targets can be, for example, voxels represented by their position and shape parameterization. Such a match creates a constraint on the trajectory point, so that a particular measurement of the environment has been acquired at different known times along the trajectory, the trajectory must be constrained in a particular way. I must. For example, assume that two measurements of the same target are made at different times, but the sensor orientation of the scanner 105 at each measurement is not accurately known. The identified voxel matches constrain the two sensor poses (relative to each other). This is because the measured targets are approximately in the same position and should be aligned. Therefore, these measurements cannot come from two sensor orientations that are aligned at will.

  Matches are determined using a k-nearest neighbor (kNN) search. A kNN search identifies the k closest points in this space given the points in the space that contain the voxel representation. The space for kNN search is a 9-dimensional space combining the position information and shape information of each voxel. Given these candidate matches, match constraints can be generated and weighted based on shape parameters (flatness, cylindricality, sphericity). Planar constraints indicate that the distance between the centroids measured along the plane normal and the difference in normal alignment should be minimized. The cylinder constraint indicates that the distance between the centroids away from the cylinder axis and the difference in axis alignment should be minimized. A spherical constraint indicates that the distance between the centroids should be minimized.

上式で、δｒ’およびδｔ’は、サンプリング時間τ_１〜τ_ｎにおける軌跡補正であり、Ａ部分行列およびｂベクトルは、３つのタイプの線形制約、すなわちマッチ制約、平滑性制約、および初期条件を表す。 At each iteration, three types of constraints can be stacked to form a sparse system with linear constraints. This sparse system can be defined by a linear system of the form

Where δr ′ and δt ′ are trajectory corrections at sampling times τ ₁ -τ _n , and the A submatrix and b vector are three types of linear constraints: match constraints, smoothness constraints, and initial conditions Represents.

  A robust statistical framework uses robustness based on Cauchy distributions to provide robustness against outliers (eg, all of the matches generated above may not be correct). The solution for the linear system is then used as a starting point for the next iteration and a new constraint is formed using the new input trajectory to generate an improved trajectory estimate. The algorithm iterates until a defined tolerance or maximum number of iterations is reached. A high level pseudo code defining this algorithm appears as Algorithm 1 in the paper cited above (see, for example, Non-Patent Document 1).

  As yet another example, an alternative registration algorithm can be described as follows. An alternative registration algorithm can iteratively refine the trajectory of the scanner 105 and the surface representing the environment to minimize projection errors in the laser measurement of the surface. Given a previous trajectory (initially from an orientation sensor such as IMU 145 and thereafter from the output of the previous iteration), all laser points are projected into a common world frame.

  A virtual 3D grid with a specific resolution (typically 0.5m x 0.5m x 0.5m) is set up and any laser points contained in the same grid cell are grouped together. An ellipsoid fits into each laser point group with sufficient membership. The centroid of the ellipsoid and the smallest eigenvector define the initial plane (surface) estimate for that grid cell. Using the error of the points in the group relative to the surface of the plane (ie, the distance between each point and the surface), the position and orientation of the plane, as well as the trajectory of the sensor head of the scanner 105, are refined.

  As described in the paper cited above (see, for example, Non-Patent Document 1), trajectory correction involves subsampling the trajectory and linearly interpolating the trajectory reference between the two temporally closest trajectory samples. To be modeled. The unknown vector (x) is a union of the trajectory subsample correction and the surface model correction. Each trajectory subsample correction has 6 degrees of freedom (3 for translation and 3 for rotation), whereas surface correction has 3 degrees of freedom, 1 for translation along the surface normal, There are two rotations about an axis perpendicular to the normal.

However, unlike the process described in the paper (eg, see Non-Patent Document 1) where each match constraint includes two trajectory references (from each matching ellipsoid), here each constraint is 1 Includes one trajectory reference and one surface reference. There is one surface constraint for each valid laser point. The smoothness constraint and the initial condition constraint can be defined in the same manner as described above. Three types of constraints can be stacked to form a linear constraint sparse system with a high level structure similar to that defined in Equation 5 above. (The difference is in the form of the A _match matrix and the contents of the x vector.)

重みｗ_ｉは、以下のように、各残差ｒ_ｉに基づく。

上式で、ｒｂａｒは、エラーに対するソフト外れ値しきい値を定義する調整可能パラメータである。Ｍ推定量フレームワークは、外れ値がある場合に制約を頑強に最適化するための標準的な統計技法である。 Trajectory and surface corrections are solved from a linearized constraint system in an iterative iterative weighted least squares M estimator framework. The surface constraints are weighted based on the Cauchy distribution of the residual r defined as

The weight w _i is based on each residual r _i as follows:

Where rbar is an adjustable parameter that defines a soft outlier threshold for error. The M estimator framework is a standard statistical technique for robustly optimizing constraints when there are outliers.

  The optimization ends when the relative change in the sum of weights is less than the threshold or when a fixed number of iterations have been completed. Full resolution trajectory corrections are interpolated from the subsampled corrections and the laser points are reprojected for subsequent iterations and final solutions.

  Referring to FIG. 5, a method for obtaining a three-dimensional scan of an environment according to one embodiment of the present invention is shown in the overall flow diagram. At step 505, the ranging device is connected to the first end of the reactive coupling mechanism. For example, a 2D laser beam scanner 105 is connected to the first end 115 of the coil spring 110.

  In step 510, the second end of the reactive linkage is connected to the object. For example, the second end 120 of the coil spring 110 is connected to the object 125.

  In step 515, the object is moved in the environment, whereby the acceleration of the object relative to the environment is converted to a movement of the distance measuring device relative to the object by a reactive coupling mechanism. This increases the field of view of the ranging device relative to the environment. For example, the translational acceleration indicated by the arrow 130 is converted into the rotational motion of the 2D laser beam scanner 105 indicated by the arrow 135 by the coil spring 110.

  In step 520, data defining a three-dimensional scan of the environment is obtained by the ranging device. For example, LIDAR data from beam 140 is collected from a surface in the environment.

  Optionally, in step 525, the computer system is connected to the ranging device. The computer system comprises a computer readable medium having stored program code components that define a registration algorithm. An example of such a computer system will be described in detail later.

  Optionally, in step 530, the data is processed using a computer system so that the registration algorithm defines the 3D point cloud of the environment.

  Referring to FIG. 6, components of a system including a computer system 600 for defining a 3D point cloud according to some embodiments of the present invention are shown in a block diagram. Computer system 600 includes a user interface 605 connected to at least one processor 610. A memory 615 is also connected to the processor 610. Memory 615 stores operating system 620, applications 625, and 3D point cloud data 630. The user interface 605 can be a combination of user interfaces including, but not limited to, a graphical user interface (GUI) such as a keypad and a computer display screen. A network interface card 640 allows the computer system 600 to connect to an electronic communication network or system 100 such as the Internet via a wired or wireless connection 645. FIG. 6 is for illustration only and includes only some components of a computer system 600 according to some embodiments of the present invention and is required for all devices capable of implementing various embodiments of the present invention. It should be understood that this is not a complete schematic representation of the various components and connections between components.

  Memory 615 also includes a computer readable program code component 635 that defines a registration algorithm. For example, when the computer readable program code component 635 is processed by the processor 610, the component 635 may perform steps of a method 500 for obtaining a 3D scan of the environment, as described above, according to some embodiments of the invention. 530 is configured to execute.

  Referring to FIG. 7, a schematic diagram of a partially cut side view of a 3D scanning beam system 700 according to an alternative embodiment of the present invention is shown. System 700 includes a ranging device in the form of a 2D laser beam scanner 705 and a reactive coupling mechanism in the form of a float 710, fluid 715, and base 720. The first end 725 (in the form of the upper end of the float 710) of the reaction linkage is connected to the 2D laser beam scanner 705. A second end 730 (in the form of a base 720) of the reaction linkage is connected to an object (not shown). A cover 735, such as an elastic membrane, can be used to both keep the fluid 715 inside the base 720 and to keep the float 710 in the center of the base 720.

  Similar to the non-fluid reactive coupling mechanism described above, according to the system 700, the translational acceleration of the object (indicated by arrow 740) is caused by rotational movement of the 2D laser beam scanner 705 (arrow 745) by the reactive coupling mechanism. Converted to Thus, a fluid coupling between the base 720 and the scanner 705 can provide an intended excitation of the movement of the scanner 705 with multiple degrees of freedom. As will be appreciated by those skilled in the art, other accelerations of the base 720, including accelerations caused by up and down forces or torsional or pivoting torques, also induce “rippling” of the fluid 715 within the base 720. Therefore, it will induce various and quasi-random 3D movements of the scanner 705. Such 3D motion of the scanner 705 and associated beam 750 can then be used to scan the environment.

  Those skilled in the art will understand that some steps of the method of the present invention can be implemented in computer program code and that the methods described herein can be implemented using various programming languages and coding implementations. It will be obvious. Further, the computer program included in the software is not limited to the specific control flow described herein, and one or more of the steps of the computer program are performed in parallel or sequentially. May be. One or more of the operations described in the context of an implementation controlled by a computer program may alternatively be implemented as a hardware electronic component.

  In summary, embodiments of the present invention include a variety of applications, including robotics, surveying, gaming, vehicle safety, object modeling, rock and soil analysis applications using underground penetrating radar, and applications of first movers. Can be used in various applications. Using the system of the present invention, functions such as map creation and / or self-location estimation, object detection, and imaging can be performed much more economically and efficiently.

  The foregoing descriptions of various embodiments of the present invention are provided for the purpose of description to those skilled in the art. This description is not intended to be exhaustive or to limit the invention to the single disclosed embodiments. As mentioned above, many alternatives and modifications to the present invention will be apparent to those skilled in the art of the above teachings. Thus, although several alternative embodiments have been specifically discussed, other embodiments will be apparent or may be relatively easily developed by those skilled in the art. Accordingly, this patent specification is intended to cover all alternatives, modifications, and variations of the invention discussed in this patent specification as well as other embodiments that fall within the spirit and scope of the invention as described above. To do.

  Limitations in the claims should be construed broadly based on the language used in the claims, and such limitations should not be limited to the specific examples described herein. . As used herein, the term “invention” is used as a reference to one or more aspects within the present disclosure. The term “invention” should not be misinterpreted as identifying critical elements, nor should it be misinterpreted as applied to all aspects and embodiments, and It should not be misinterpreted as limiting.

Claims (20)

A ranging device;
A reaction coupling mechanism having a first end and a second end;
The first end is connected to the ranging device;
The second end is connected to an object that moves the system in an environment;
In use, a three-dimensional scanning beam system in which acceleration of the object relative to the environment is converted into motion of the distance measuring device relative to the object by the reactive coupling mechanism to increase the field of view of the distance measuring device relative to the environment .   The system of claim 1, further comprising an orientation sensor attached to the ranging device.   A computer system connected to the ranging device, the computer comprising a computer readable medium storing a program code component defining a registration algorithm, wherein the registration algorithm receives data from the ranging device; The system of claim 1, processing to define a 3D point cloud of the environment.   The system of claim 1, wherein the ranging device comprises a laser.   The system of claim 4, wherein the laser comprises a two-dimensional scanning beam laser.   The system of claim 3, wherein the 3D point cloud is defined in real time.   The system of claim 1, wherein the ranging device transmits and receives at least one of the following types of signals: optical signal, acoustic signal, ultrasonic signal, radio frequency signal, γ radiation signal, microwave signal.   The system of claim 1, wherein the reactive coupling mechanism comprises a spring.   The system of claim 2, wherein the orientation sensor comprises an inertial measurement unit (IMU).   The system of claim 1, wherein the connection between the first end of the reactive coupling mechanism and the ranging device is a fluid coupling. A method for obtaining a three-dimensional scan of an environment,
Connecting the ranging device to the first end of the reactive coupling mechanism;
Connecting the second end of the reactive linkage to an object;
Moving the object in the environment, wherein the acceleration of the object is converted into movement of the distance measuring device relative to the object by the reactive coupling mechanism, increasing the field of view of the distance measuring device relative to the environment. And
Obtaining data defining the three-dimensional scan of the environment by the ranging device.   The method according to claim 11, wherein an orientation sensor is attached to the distance measuring device. Connecting a computer system comprising a computer readable medium storing a program code component defining a registration algorithm to the ranging device; and
The method of claim 11, further comprising processing the data using the computer system, wherein the registration algorithm further defines a 3D point cloud of the environment.   The method of claim 11, wherein the ranging device comprises a laser.   The method of claim 14, wherein the laser comprises a two-dimensional scanning beam laser.   The method of claim 13, wherein the 3D point cloud is defined in real time.   The method according to claim 11, wherein the ranging device transmits and receives at least one of the following types of signals: optical signal, acoustic signal, ultrasonic signal, radio frequency signal, γ radiation signal, microwave signal.   The method of claim 11, wherein the reactive coupling mechanism comprises a spring.   The method of claim 12, wherein the orientation sensor comprises an inertial measurement unit (IMU).   The method of claim 11, wherein the reactive coupling mechanism comprises a flowable bond.

Images