Measuring device for rapidly measuring horizontal reference of desktop

Abstract

The invention relates to the technical field of measurement, in particular to a desktop horizontal reference fast measurement device which is used for carrying out horizontal reference and flatness measurement on a handheld scanned plane to be measured, and a processing unit is used for obtaining a depth sensing unit of spatial point depth information data flow of a local area of the plane to be measured. The inertial measurement unit is used for acquiring a motion state of the measurement device and a data stream reflecting attitude information of the measurement device relative to a gravity direction, and the processing unit is connected with the depth sensing unit and the inertial measurement unit; and estimating the pose of the measuring device based on the depth information data stream and the inertial measurement unit, constructing a three-dimensional point cloud map of the plane to be measured, executing global level calibration based on the constructed three-dimensional point cloud map, and executing geometric analysis on the calibrated three-dimensional point cloud map to determine the overall levelness and local flatness of the plane to be measured. According to the invention, rapid three-dimensional modeling and accurate levelness and flatness analysis of the plane to be measured are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of measurement, and it relates to a measurement device for quickly measuring the desktop horizontal reference. Background Art

[0002] In the fields of modern architectural decoration, furniture installation, industrial manufacturing, etc., the levelness and flatness of plane structures such as desktops and countertops are important indicators to measure their installation quality and service performance. Currently, the main tools used to detect whether such planes are horizontal or flat include bubble levels, laser line projectors, and digital levels based on electronic sensors and other devices. However, these traditional measurement methods and devices have many limitations in the actual application process.

[0003] The traditional bubble level relies on the position change of bubbles in liquid alcohol to judge whether the surface is horizontal. Its structure is simple and the cost is low, but it can only reflect the inclination state in a single direction, cannot detect local unevenness, and has low accuracy, making it difficult to meet the requirements of fine construction.

[0004] The laser level sets the horizontal reference by emitting visible laser lines. Although it has the advantages of strong intuitiveness and wide application range, when detecting non-regular surfaces or desktops with local undulations, it can only provide reference lines and lacks the ability to comprehensively evaluate the surface topography, and is more sensitive to external light conditions, which limits its application in complex environments.

[0005] In recent years, the gradually popularized electronic digital level uses inertial sensors such as accelerometers or gyroscopes to achieve digital measurement and can provide high-precision angle measurement data. However, this type of device essentially still belongs to a single-point or multi-point angle measurement system, which can only reflect the overall inclination and cannot sense the surface height difference at the microscopic level, so it is difficult to accurately judge the local flatness of the desktop.

[0006] In addition, the existing technologies generally lack the ability of intelligent analysis. Users need to interpret the measurement results by themselves and make judgments, which is a relatively high threshold for non-professionals. At the same time, most devices do not have a graphical feedback function and cannot intuitively display the state of the entire surface to be measured, affecting the use experience and work efficiency.

[0007] In summary, the desktop level detection devices on the current market generally have problems such as single detection dimension, poor adaptability, low intelligence level, and weak visualization effect. There is an urgent need for a new type of detection device with novel structure, unique principle, comprehensive functions, and convenient operation to achieve the synchronous and rapid detection of the overall level and local flatness of the desktop. Summary of the Invention

[0008] To this end, the object of the present invention is to provide a measurement device for quickly measuring the horizontal reference of a desktop. By fusing depth perception, inertial measurement, and optional visual perception data, and using advanced Simultaneous Localization and Mapping (SLAM) technology and global horizontal calibration based on the gravity vector, rapid three-dimensional modeling of the plane to be measured and accurate horizontal and flatness analysis are achieved.

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] A measurement device for quickly measuring the horizontal reference and flatness of a plane to be measured by handheld scanning, comprising:

[0011] A depth perception unit for obtaining a data stream of the depth information of spatial points in a local area of the plane to be measured during the movement of the user holding the measurement device;

[0012] An inertial measurement unit for obtaining a data stream of the motion state of the measurement device and the attitude information reflecting the measurement device's orientation relative to the direction of gravity;

[0013] A processing unit connected to the depth perception unit and the inertial measurement unit, configured to:

[0014] Execute a Simultaneous Localization and Mapping algorithm based on the depth information data stream and the motion state and attitude information data stream of the inertial measurement unit to estimate the pose of the measurement device and construct a three-dimensional point cloud map of the plane to be measured;

[0015] Perform global horizontal calibration on the constructed three-dimensional point cloud map based on the attitude information reflecting the orientation relative to the direction of gravity provided by the inertial measurement unit, so that the calibrated point cloud map is aligned with the true horizontal reference plane; and

[0016] Perform geometric analysis on the calibrated three-dimensional point cloud map to determine the overall levelness and local flatness of the plane to be measured.

[0017] The present invention is further configured such that the processing unit is configured to perform non-linear optimization based on a multi-sensor data joint model in the Simultaneous Localization and Mapping algorithm, and the model includes at least one measurement residual term based on the readings of the inertial measurement unit, and the calculation of this residual term is related to the gravity vector.

[0018] The present invention is further configured such that the non-linear optimization based on the multi-sensor data joint model aims to minimize the total error including the measurement residuals of the inertial measurement unit and the depth perception unit. The construction of the measurement residual term of the inertial measurement unit takes into account the motion state, attitude changes of the measurement device, and the influence of the gravity vector.

[0019] The present invention is further configured such that the measuring device further includes a visual perception unit connected to the processing unit for acquiring an image data stream of the plane to be measured, and the processing unit is configured to fuse the image data stream in a simultaneous localization and mapping algorithm to estimate the pose of the measuring device.

[0020] The present invention is further configured such that the processing unit is configured to, when performing global horizontal calibration, determine the rotation relationship of the coordinate system of the three-dimensional point cloud map relative to the true horizontal reference plane based on the attitude estimation of the measuring device relative to the gravity direction obtained from the inertial measurement unit information, and apply this rotation relationship to transform the point cloud map.

[0021] The present invention is further configured such that the processing unit is configured to, when performing geometric analysis, perform plane fitting on the calibrated three-dimensional point cloud map and determine the overall levelness based on the relationship between the fitted plane and the true horizontal reference plane.

[0022] The present invention is further configured such that the processing unit is configured to, when performing geometric analysis, calculate the vertical deviation of the points in the calibrated point cloud map relative to the plane fitting result and calculate the local flatness based on the vertical deviation.

[0023] The present invention is further configured such that the processing unit is configured to calculate the difference between the maximum value and the minimum value of the vertical deviation as the peak-to-valley difference index of the local flatness, and calculate the root mean square of the vertical deviation as the root mean square deviation index of the local flatness.

[0024] The present invention is further configured such that the measuring device includes a communication unit connected to the processing unit for wirelessly transmitting the measurement results and the three-dimensional point cloud map to an external computing unit, and the external computing unit is configured to further analyze or display the results or the map.

[0025] The present invention is further configured such that the measuring device includes a housing for integrating the perception unit, the processing unit, and the communication unit, and the design of the housing facilitates the user to hold the measuring device and move it above the plane to be measured for scanning.

[0026] Compared with the deficiencies of the prior art, the beneficial effects of the present invention are as follows:

[0027] By fusing IMU data and performing global horizontal calibration based on the gravity vector, this device measures the direction of the plane to be measured relative to the true physical horizontal plane, rather than just relative to the starting position or the scanning path. This provides a levelness evaluation with physical significance and high precision.

[0028] It not only provides the overall levelness (tilt angle), but also gives detailed local flatness quantification indexes within the entire scanned area.

[0029] Generate a dense 3D point cloud map, which can be used for visualization, export, and further detailed analysis, archiving, or quality control records, going beyond simple numerical readings. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic diagram of the framework of the present invention;

[0031] Figure 2 It is a schematic flowchart of the processing unit of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] Refer to Figures 1 to 2 A measurement device for quickly measuring the desktop horizontal reference of the present invention is further described. It includes: a housing, a depth sensing unit integrated in the housing, an inertial measurement unit, a processing unit, and a communication unit.

[0033] The housing is made of lightweight and high-strength engineering plastics (such as ABS or PC). Through ergonomic design, it ensures comfortable holding, stable center of gravity, and is convenient for users to perform smooth scanning with one hand. The internal structure design of the housing needs to ensure the relative positions of the sensor units are fixed to ensure the effectiveness of the calibration parameters. It is provided with a power switch, status indicators (such as power, scanning, completion, error), and a possible USB-C interface (for charging or wired data transmission).

[0034] Depth sensing unit: A depth camera using active structured light or time-of-flight (ToF) technology.

[0035] Parameters: Outputs depth images and infrared images (or color images if integrated) with a resolution of 640x480, a frame rate of 30FPS, an effective ranging range of 0.2m - 3m, and an accuracy of millimeter level.

[0036] During the scanning process, the three-dimensional spatial point coordinates (relative to the camera coordinate system) of the local area of the plane to be measured within the field of view of the measurement device are obtained in real time to form a depth information data stream

[0037] , where .

[0038] The depth information data stream is a series of timestamped depth maps continuously captured by the depth sensing unit at a certain frame rate during the user's hand-held device scanning process. Each depth map contains the three-dimensional distance information of the local area of the plane to be measured within the sensor's field of view at that moment. This continuous data stream is one of the basic inputs for the subsequent SLAM algorithm to construct a complete three-dimensional point cloud map.

[0039] Inertial Measurement Unit (IMU): A six-axis MEMS IMU integrating a three-axis accelerometer and a three-axis gyroscope. For example, Bosch BMI088 or TDK InvenSense MPU-6050. Hardware-synchronized with the depth perception unit or with precise timestamps.

[0040] Parameters: Accelerometer range ±8g, gyroscope range ±2000dps, data output rate 200Hz or higher. Low noise density and good bias stability.

[0041] Functions: Real-time measurement of linear acceleration in the device body coordinate system and angular velocity data stream. By the accelerometer readings, the direction of the gravity vector can be sensed under static or quasi-static conditions, thus providing attitude information (such as pitch angle and roll angle) of the device relative to the earth (gravity) coordinate system.

[0042] Processing Unit (PU): A high-performance and low-power embedded computing platform. At least 4GB of RAM, with sufficient computing power (CPU + GPU) to run the SLAM algorithm and subsequent processing in real time. Runs the embedded Linux operating system and custom-developed measurement software. Connects to the DPU, IMU, and VPU via MIPI CSI, USB, or an internal bus.

[0043] Communication Unit (CU): Consists of Wi-Fi (802.11ac) and / or Bluetooth (BLE5.0) modules.

[0044] Functions: Wirelessly transmits the calculated levelness / flatness results and the calibrated 3D point cloud map (such as in.ply or.pcd format) to a smartphone App, tablet, or PC software for visual display, report generation, or further analysis.

[0045] Power Supply Unit: Built-in rechargeable lithium-ion battery (e.g., 3000mAh), providing at least 1 hour of continuous working time. Includes a power management circuit.

[0046] Core Algorithms and Working Principles:

[0047] After the device is started, data acquisition begins synchronously. The data stream has precise timestamps and is sent to the processing unit. The IMU data is initially filtered and denoised, and the depth image needs to be denoised and invalid points removed.

[0048] Simultaneous Localization and Mapping (SLAM): Adopts a tightly coupled visual-inertial-depth SLAM framework. The estimated state vector X contains at least the pose (position and attitude, and the attitude is usually represented by a quaternion or a rotation matrix Indicates that, where W is the world coordinate system and B is the IMU coordinate system of the device body), velocity , and the accelerometer bias of the IMU and the gyroscope bias .

[0049] ,

[0050] where is the state of the th key frame, is an optional set of 3D map points (landmarks).

[0051] Factor graph construction and optimization:

[0052] IMU pre-integration factor: Connects the states of two consecutive key frames and +1 . Using all IMU measurements between two key frames , calculates the relative motion constraint and its covariance through the IMU pre-integration theory. The IMU measurement model is:

[0053]

[0054]

[0055] where is the true acceleration in the world coordinate system, is the gravity vector in the world coordinate system (usually set to , or optimized estimate), is the rotation from the world to the body, is the body angular velocity, , is the noise.

[0056] The measurement residual term of pre-integration is used to constrain adjacent states in the optimization. The key is that the calculation of this residual explicitly or implicitly depends on the gravity vector . For example, the gravity influence needs to be subtracted in the state transition equation, or the gravity direction is used as part of the variables to be estimated in the optimization. The optimization objective is to minimize the squared Mahalanobis norm of this residual: .

[0057] Depth / vision factor: For each key frame k, the observed 3D points (from DPU) or 2D feature points (from VPU) . If the point already exists in the map , then construct the reprojection error factor.

[0058] For the depth points (camera coordinate system), transform to the world coordinate system ( is the homogeneous transformation matrix, is the extrinsic parameter from the IMU to the camera).

[0059] For the visual feature points (pixel coordinates), the corresponding 3D map points are back-projected onto the image plane of the current frame k to obtain the predicted pixel coordinates ( is the camera projection model).

[0060] The residual term or calculates the difference between the measured value and the predicted value. For example , the optimization objective is to minimize .

[0061] Joint optimization: The optimization objective is to minimize the weighted sum of squares of all factor residuals (Mahalanobis distance based on the covariance matrix ):

[0062]

[0063] Iteratively solve for the optimal state X^ using a non-linear least squares solver (such as Gauss-Newton). This joint optimization framework ensures that the motion and gravity information provided by the IMU and the geometric information provided by depth / vision can constrain and correct each other, resulting in a more accurate pose estimation and map.

[0064] Map construction: Transform the depth points observed in each key frame to the unified world coordinate system

[0065] according to the optimized pose of that frame and cumulatively construct a dense 3D point cloud map of the plane to be measured .

[0066] Global horizontal calibration: Rotate the point cloud map constructed by SLAM (the direction of the Z axis of its coordinate system W is determined by SLAM initialization and is usually not aligned with the gravity direction) to a new coordinate system such that the XY plane of is the true horizontal plane and the Z axis is perpendicular to the horizontal plane (i.e., aligned in the opposite direction to the gravity vector).

[0067] During the SLAM optimization process, the acceleration readings of the IMU (especially at low dynamics) provide information about the gravity vector Estimation in the world coordinate system W. The optimized IMU attitude and bias can be utilized. Combining with the original accelerometer readings, estimate at multiple moments .

[0068] Calculate the rotation relationship: It is necessary to find a rotation matrix such that , (assuming the positive direction of the Z-axis in the H system is upward).

[0069] Let the target vector be ; calculate the unit vector of ;

[0070] Calculate the rotation axis ; calculate the rotation angle .

[0071] According to the axis-angle calculate the rotation matrix (for example, using the Rodriguez rotation formula).

[0072] Apply the transformation: For each point in the point cloud map perform a rotation transformation:

[0073]

[0074] to obtain the calibrated point cloud map . At this time the point coordinates in are referenced to the true horizontal plane.

[0075] Geometric analysis: Input the calibrated three-dimensional point cloud map ;

[0076] Adopt the RANSAC (Random Sample Consensus) algorithm. Randomly select the minimum point set (3 points) to fit a plane model ax + by + cz + d = 0, and then count the number of points (inliers) in the point cloud whose distance to this plane is within the threshold. Repeat this process and select the plane with the most inliers as the best-fitting plane . RANSAC can effectively resist the noise points or outliers that may be generated during the scanning process.

[0077] Output the parameters (a, b, c, d) of the fitted plane and the inlier set .

[0078] Overall levelness calculation: The normal vector of the fitted plane .

[0079] The normal vector of the true horizontal plane (in the H system) is (or to maintain direction consistency).

[0080] Levelness index: Calculate the angle between two normal vectors :

[0081] The smaller this angle, the closer the overall plane to be measured is to horizontal. The result can be reported in degrees or radians.

[0082] Local flatness calculation:

[0083] Calculate the vertical deviation: For each point in the inlier set of the fitted plane , calculate its vertical distance to the fitted plane (i.e., the deviation along the Z-axis direction, since the plane is approximately horizontal):

[0084]

[0085] (assuming , if then the plane is approximately vertical and not applicable to levelness measurement). More precisely, it is the orthogonal distance from the point to the plane: ; but after horizontal calibration, usually focus on deviation.

[0086] Flatness index:

[0087] Peak-to-valley difference: Calculate the difference between the maximum and minimum values of all inlier vertical deviations:

[0088]

[0089] Root mean square deviation: Calculate the root mean square value of all inlier vertical deviations:

[0090]

[0091] The smaller these two indices, the smaller the surface undulation of the plane to be measured, that is, the flatter it is.

[0092] Result output and transmission: The processing unit calculates numerical results such as

[0093] Through the communication unit (CU), wirelessly transmit these numerical results and (optional) calibrated point cloud data to a paired external computing device (such as a mobile phone App).

[0094] The software on the external device is responsible for receiving data, performing visual display (such as displaying the level angle, flatness value, rendering a three-dimensional point cloud map and using color mapping to show the deviation magnitude, etc.), and generating a measurement report.

[0095] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any ordinary changes and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included in the protection scope of the present invention.

Claims

1. A measuring device for quickly measuring the desktop horizontal reference, which is used to measure the horizontal reference and flatness of the plane to be measured by handheld scanning, and is characterized in that, Comprising: A depth perception unit, configured to acquire a data stream of spatial point depth information of a local area of the plane to be measured during the movement of the user holding the measurement device; An inertial measurement unit, configured to acquire a data stream of the motion state of the measurement device and attitude information reflecting the attitude of the measurement device relative to the direction of gravity; A processing unit, connected to the depth perception unit and the inertial measurement unit, and configured as follows: Based on the data stream of depth information and the motion state and attitude information data stream of the inertial measurement unit, execute a simultaneous localization and mapping algorithm to estimate the pose of the measurement device and construct a three-dimensional point cloud map of the plane to be measured; Based on the attitude information reflecting the direction relative to gravity provided by the inertial measurement unit, perform global horizontal calibration on the constructed three-dimensional point cloud map, so that the calibrated point cloud map is aligned with the true horizontal reference plane; And Perform geometric analysis on the calibrated three-dimensional point cloud map to determine the overall levelness and local flatness of the plane to be measured.

2. The measuring device for quickly measuring the desktop horizontal reference according to claim 1, characterized in that, The processing unit is configured to execute a non-linear optimization based on a multi-sensor data joint model in the simultaneous localization and mapping algorithm. The model includes at least one measurement residual term based on the readings of the inertial measurement unit, and the calculation of this residual term is related to the gravity vector.

3. The measuring device for quickly measuring the desktop horizontal reference according to claim 2, characterized in that, The non-linear optimization based on the multi-sensor data joint model aims to minimize the total error including the measurement residual of the inertial measurement unit and the measurement residual of the depth perception unit. The construction of the measurement residual term of the inertial measurement unit takes into account the motion state, attitude change of the measurement device, and the influence of the gravity vector.

4. The measuring device for quickly measuring the desktop horizontal reference according to claim 1, characterized in that, The measurement device further includes a visual perception unit, connected to the processing unit, for acquiring an image data stream of the plane to be measured. The processing unit is configured to fuse the image data stream in the simultaneous localization and mapping algorithm to estimate the pose of the measurement device.

5. A measuring device for quickly measuring the desktop horizontal reference according to claim 1, characterized in that, The processing unit is configured to, when performing global horizontal calibration, determine the rotation relationship between the coordinate system of the three-dimensional point cloud map and the true horizontal reference plane based on the attitude estimation of the measurement device relative to the direction of gravity obtained from the inertial measurement unit information, and apply this rotation relationship to transform the point cloud map.

6. The measuring device for quickly measuring the desktop horizontal reference according to claim 1, characterized in that, The processing unit is configured to, when performing geometric analysis, perform plane fitting on the calibrated three-dimensional point cloud map, and determine the overall levelness based on the relationship between the fitted plane and the true horizontal reference plane.

7. The measuring device for quickly measuring the desktop horizontal reference according to claim 6, wherein, The processing unit is configured to, when performing geometric analysis, calculate the vertical deviation of the points in the calibrated point cloud map relative to the plane fitting result, and calculate the local flatness based on the vertical deviation.

8. The measuring device for quickly measuring the desktop horizontal reference according to claim 7, characterized in that, The processing unit is configured to calculate the difference between the maximum value and the minimum value of the vertical deviation as the peak-valley difference index of the local flatness, and calculate the root mean square of the vertical deviation as the root mean square deviation index of the local flatness.

9. The measuring device for quickly measuring the desktop horizontal reference according to claim 1, characterized in that, The measurement device includes a communication unit, connected to the processing unit, for wirelessly transmitting the measurement results and the three-dimensional point cloud map to an external computing unit, and the external computing unit is configured to further analyze or display the results or the map.

10. The measuring device for quickly measuring the desktop horizontal reference according to claim 1, characterized in that, The measurement device includes a housing, used to integrate the sensing unit, the processing unit, and the communication unit. The design of the housing facilitates the user to hold the measurement device and move it above the plane to be measured for scanning.

Images