TW202137138A

TW202137138A - Method and system of generating a geodetic coordinates 3d point cloud map

Info

Publication number: TW202137138A
Application number: TW109108869A
Authority: TW
Inventors: 彭正偉; 楊承翰
Original assignee: 勤崴國際科技股份有限公司
Priority date: 2020-03-18
Filing date: 2020-03-18
Publication date: 2021-10-01

Abstract

The present invention provides a method and a system of generating a Geodetic coordinates 3D point cloud map. In order to correct 3D point cloud, a number of pre-survey control reference points are applied to generate a corrected 3D point cloud. The corrected 3D point cloud helps Lidar localization method to work out a precise trajectory. It will compensate the position error of position data from satellite positioning system. Then, a plurality of 3D point clouds can be reconstruct and coincided into a 3D point cloud map with accurate position.

Description

含大地座標之3D點雲地圖的產生方法和3D點雲地圖產生系統Method for generating 3D point cloud map containing geodetic coordinates and system for generating 3D point cloud map

本發明涉及地圖資料處理領域，具體而言涉及3D點雲地圖的產生方法和系統。The invention relates to the field of map data processing, in particular to a method and system for generating a 3D point cloud map.

光學雷達藉由其中的感測器發射而經物體反射回接收器的雷射脈衝，並測量雷射脈衝的來回時間，進而能夠精準判斷感測器和物體之間的距離。由於光學雷達發出及接收雷射脈衝可達相當高頻率，甚至每秒可高達數萬至數十萬雷射脈衝，且可透過增加發射／接收組取得對於感知環境的輪廓，因此目前光學雷達已普遍用於行車感知輔助相關的應用。然而，光學雷達藉接收雷射脈衝產生具空間位置相關資訊的點雲尚需要進一步處理來降低誤差。The optical radar uses the laser pulse emitted by the sensor and reflected by the object back to the receiver, and measures the round-trip time of the laser pulse, so as to accurately determine the distance between the sensor and the object. Since the optical radar sends and receives laser pulses up to a very high frequency, even up to tens of thousands to hundreds of thousands of laser pulses per second, and the profile of the sensing environment can be obtained by increasing the transmitting/receiving group, the current optical radar has It is generally used in applications related to driving perception assistance. However, the optical radar generates a point cloud with spatial position-related information by receiving laser pulses and needs further processing to reduce errors.

本發明之一目的在於提供一種含大地座標之3D點雲地圖的產生方法與系統，可給定多個3D點雲中特徵點的三維座標從而校正3D點雲，並根據該已被校正的3D點雲結合光達定位技術將有效修正衛星定位系統之定位誤差，從而製成含大地座標之高精度座標之3D點雲地圖。An object of the present invention is to provide a method and system for generating a 3D point cloud map containing geodetic coordinates, which can be given the three-dimensional coordinates of feature points in a plurality of 3D point clouds to correct the 3D point cloud, and based on the corrected 3D point cloud. The point cloud combined with LiDAR positioning technology will effectively correct the positioning error of the satellite positioning system, thereby creating a 3D point cloud map with high-precision coordinates containing geodetic coordinates.

依據本發明一實施例，提供一種含大地座標之3D點雲地圖的產生方法，應用於包括多個特徵點的多筆3D點雲，包括：產生一經校正的3D點雲，根據該經校正的3D點雲，修正由一定位系統提供的定位資料之誤差，及產生一含大地座標之3D點雲地圖。According to an embodiment of the present invention, a method for generating a 3D point cloud map containing geodetic coordinates is provided, which is applied to a multi-stroke 3D point cloud including a plurality of feature points, including: generating a corrected 3D point cloud based on the corrected The 3D point cloud corrects the errors of the positioning data provided by a positioning system and generates a 3D point cloud map with geodetic coordinates.

依據本發明一實施例，產生一經校正的3D點雲的步驟包括：對複數個3D點雲分別進行下列步驟：取得3D點雲內一特徵點並給定各特徵點一精準量測的控制點之三維大地座標，依照對應該特徵點給定的控制點之三維大地座標之間的相對位置，對所有特徵點間的取樣點位置做比例的水平、垂直移動校正。這些3D點雲中在特徵點間的多個取樣點即被控制點的絕對位置有效修正。According to an embodiment of the present invention, the step of generating a calibrated 3D point cloud includes: performing the following steps on a plurality of 3D point clouds: obtaining a feature point in the 3D point cloud and assigning each feature point a precisely measured control point For the three-dimensional geodetic coordinates, according to the relative position between the three-dimensional geodetic coordinates of the control points given corresponding to the feature points, the proportional horizontal and vertical movement corrections are made to the sampling point positions between all feature points. The multiple sampling points between the feature points in these 3D point clouds are effectively corrected by the absolute positions of the control points.

依據本發明一實施例，產生一經校正的3D點雲的步驟包括：判斷對應該控制點的一預定範圍內要校正的多個取樣點，依照對應該特徵的該控制點與該取樣點的相對位置校正該預定範圍內要校正的該些取樣點，及判斷經校正的該些3D點雲的該些取樣點是否符合精度要求。According to an embodiment of the present invention, the step of generating a corrected 3D point cloud includes: judging a plurality of sampling points to be corrected within a predetermined range corresponding to the control point, according to the relative characteristics of the control point and the sampling point The position is corrected for the sampling points to be corrected within the predetermined range, and it is determined whether the sampling points of the corrected 3D point clouds meet the accuracy requirements.

依據本發明一實施例，根據該經校正的3D點雲，修正由一定位系統提供的定位資料之誤差之步驟包括：依據該經校正的3D點雲取得一光達定位結果，比較該光達定位結果與該定位資料的誤差，修正該定位資料。According to an embodiment of the present invention, the step of correcting errors in positioning data provided by a positioning system according to the corrected 3D point cloud includes: obtaining a LiDAR positioning result based on the corrected 3D point cloud, and comparing the LiDAR positioning result The error between the positioning result and the positioning data is corrected.

依據本發明一實施例，根據該經校正的3D點雲，修正由一定位系統提供的定位資料之誤差之步驟包括：判斷超過一預定誤差範圍時，修正該定位資料，且判斷不超過該預定誤差範圍時，不修正該定位資料；產出重新建構之該修正後精確定位軌跡。According to an embodiment of the present invention, according to the corrected 3D point cloud, the step of correcting the error of the positioning data provided by a positioning system includes: when it is determined that the error range exceeds a predetermined error range, the positioning data is corrected, and it is determined that the predetermined error range is not exceeded. When the error is within the range, the positioning data is not corrected; the reconstructed accurate positioning trajectory after the correction is generated.

依據本發明一實施例，該光達定位結果係由一光學雷達的迭代最近點（Iterative Closest Point，簡稱ICP）、迭代最近投影點（Iterative Closest Projected Point，簡稱ICPP）、正態分佈變換（Normal Distribution Transform，簡稱NDT）之一計算方式所產生。According to an embodiment of the present invention, the LiDAR positioning result is based on an optical radar's Iterative Closest Point (ICP), Iterative Closest Projected Point (ICPP), and Normal Distribution Transformation (Normal). Distribution Transform, referred to as NDT) is generated by one of the calculation methods.

依據本發明另一實施例，提供一種3D點雲地圖產生系統，包括：一移動式測繪平台，其中裝設有一導航定位裝置、一光學雷達及一資料處理單元，其中該導航定位裝置接收來自一定位系統提供的定位資料，該資料處理單元可接收來自導航定位裝置的定位資料及來自光學雷達的多筆3D點雲，每一筆3D點雲包括多個取樣點；及一資料後製系統，接收該些3D點雲及至少一控制點的位置，且該資料後製系統係被配置以執行以下步驟：產生一經校正的3D點雲，根據該經校正的3D點雲，修正該定位資料之誤差，及產生一3D點雲地圖。According to another embodiment of the present invention, a 3D point cloud map generation system is provided, including: a mobile surveying and mapping platform, in which a navigation positioning device, an optical radar, and a data processing unit are installed, wherein the navigation positioning device receives data from a The positioning data provided by the positioning system, the data processing unit can receive positioning data from the navigation positioning device and multiple 3D point clouds from the optical radar, each 3D point cloud including multiple sampling points; and a data post-processing system to receive The positions of the 3D point clouds and at least one control point, and the data post-processing system is configured to perform the following steps: generate a corrected 3D point cloud, and correct the error of the positioning data based on the corrected 3D point cloud , And generate a 3D point cloud map.

依據本發明一實施例，該導航定位裝置包括一全球定位系統接收器及一慣性量測單元，全球定位系統接收器接收來自衛星之全球定位系統定位資料，慣性量測單元連續估算即時位置，定位資料為全球定位系統/慣性導航系統定位資料。According to an embodiment of the present invention, the navigation and positioning device includes a GPS receiver and an inertial measurement unit. The GPS receiver receives GPS positioning data from satellites. The inertial measurement unit continuously estimates the real-time position and positions The data is GPS/Inertial Navigation System positioning data.

因此，本發明3D點雲地圖的產生方法和系統，可透過控制點直接校正特徵點並水平、垂直依距離比例移動取樣點而產生經校正的3D點雲，並依據經校正的3D點雲修正定位資料的誤差，從而提升三維位置上的精確度，並將3D點雲疊合而產生具有高擬合、高精度位置的3D點雲地圖。Therefore, the method and system for generating a 3D point cloud map of the present invention can directly correct the characteristic points through the control points and move the sampling points horizontally and vertically according to the distance ratio to generate a corrected 3D point cloud, and correct it according to the corrected 3D point cloud The error of the positioning data is improved to improve the accuracy of the three-dimensional position, and the 3D point clouds are superimposed to produce a 3D point cloud map with high fitting and high-precision position.

圖1繪示本發明一實施例所提供的含大地座標之三維（3D）點雲地圖產生系統10之方塊圖。含大地座標之3D點雲地圖產生系統10包括一移動式測繪平台100及一資料後製系統200。移動式測繪平台100例如為車載之測繪平台，在其行車中可就環境擷取不同類型之初始資料122，舉例來說，初始資料122包括測繪平台移動時的軌跡及光達在每個時間點所收集3D點雲的相對位置訊息。隨後，初始資料122可由資料後製系統200加以處理，包括校正取樣點的三維座標而接合3D點雲並修正一定位系統之定位資料（在此可包含一全球定位系統（Global Positioning System, GPS）提供之定位資料，以下稱「GPS定位資料」）的誤差，從而製成高精度座標之3D點雲地圖。FIG. 1 shows a block diagram of a system 10 for generating a three-dimensional (3D) point cloud map containing geodetic coordinates according to an embodiment of the present invention. The 3D point cloud map generating system 10 containing geodetic coordinates includes a mobile surveying and mapping platform 100 and a data post-production system 200. The mobile surveying and mapping platform 100 is, for example, a vehicle-mounted surveying and mapping platform. Different types of initial data 122 can be captured in the driving environment. For example, the initial data 122 includes the trajectory of the surveying and mapping platform when it moves and the time point of Lidar. The relative position information of the collected 3D point cloud. Subsequently, the initial data 122 can be processed by the data post-processing system 200, including correcting the three-dimensional coordinates of the sampling points, joining the 3D point cloud, and correcting the positioning data of a positioning system (which may include a Global Positioning System (GPS)) The provided positioning data, hereinafter referred to as "GPS positioning data") error, to create a 3D point cloud map with high-precision coordinates.

根據一實施例，移動式測繪平台100可包括一導航定位裝置102、一光學雷達104及一資料處理單元110。根據一實施例，導航定位裝置102例如包括GPS接收器，其可接收來自衛星之GPS定位資料和時間資料，將之提供給時間同步模組112及光學雷達104。GPS定位資料可以是一組定位定向解或其他形式，在此無須限制。根據一實施例，導航定位裝置102尚整合慣性量測單元(Inertial Measurement Unit)利用其加速計、陀螺儀和計算機來連續估算即時位置達到高可靠度INS。According to an embodiment, the mobile surveying and mapping platform 100 may include a navigation and positioning device 102, an optical radar 104, and a data processing unit 110. According to an embodiment, the navigation and positioning device 102 includes, for example, a GPS receiver, which can receive GPS positioning data and time data from satellites, and provide them to the time synchronization module 112 and the optical radar 104. GPS positioning data can be a set of positioning solutions or other forms, and there is no restriction here. According to an embodiment, the navigation and positioning device 102 further integrates an inertial measurement unit (Inertial Measurement Unit) using its accelerometer, gyroscope, and computer to continuously estimate the real-time position to achieve a highly reliable INS.

光學雷達104可包括至少一組光達發射器與光達接收器，光達接收器以一定頻率接收光達發射器發出的雷射脈衝，依據發射角度及雷射信號飛行時間資訊推算出取樣點的3D位置資訊，而多個角度所發射出的多個取樣點最後可推算出物件輪廓。由光學雷達104提供的3D點雲中的各取樣點之時間資訊與3D位置資訊，例如包括取樣點的收集時間及基於導航定位裝置102位置資訊及姿態資訊，計算出之反射物體位置資訊。The optical radar 104 may include at least one set of LiDAR transmitter and LiDAR receiver. The LiDAR receiver receives laser pulses from the LiDAR transmitter at a certain frequency, and calculates the sampling point based on the launch angle and the flight time information of the laser signal. 3D position information, and multiple sampling points emitted from multiple angles can finally calculate the contour of the object. The time information and 3D position information of each sampling point in the 3D point cloud provided by the optical radar 104 include, for example, the collection time of the sampling point and the position information of the reflective object calculated based on the position information and attitude information of the navigation positioning device 102.

資料處理單元110例如為工業電腦、車用電腦等。資料處理單元110可接收來自導航定位裝置102及光學雷達104之資料，進行資料格式轉換或壓縮等資料預處理步驟。根據一實施例，資料處理單元110中所配置的軟體架構可包括一時間同步模組112、一資料紀錄模組114和一作業系統116。作業系統116可管理硬體和軟體之資源、為軟體執行基本服務與外界的連線。時間同步模組112可根據導航定位裝置102所提供的資訊更新作業系統116的時鐘，使其與導航定位裝置102時間同步，進而與光學雷達104產生的點資料時間同步。資料紀錄模組114可產生檔案紀錄資料。The data processing unit 110 is, for example, an industrial computer, a car computer, or the like. The data processing unit 110 can receive data from the navigation and positioning device 102 and the optical radar 104, and perform data preprocessing steps such as data format conversion or compression. According to an embodiment, the software architecture configured in the data processing unit 110 may include a time synchronization module 112, a data recording module 114, and an operating system 116. The operating system 116 can manage hardware and software resources, and perform basic services for the software to connect to the outside world. The time synchronization module 112 can update the clock of the operating system 116 according to the information provided by the navigation and positioning device 102 to synchronize it with the time of the navigation and positioning device 102, and then to synchronize the time with the point data generated by the optical radar 104. The data record module 114 can generate file record data.

資料後製系統200例如為個人電腦、工業電腦等。資料後製系統200可取得來自資料處理單元110的初始資料122並接收輸入之控制點的位置124及檢核點的位置126，初始資料122包括但不限於，由光學雷達104所取得的包括多個取樣點的多筆3D點雲。資料後製系統200可透過任何合適的方式取得初始資料122，例如：資料後製系統200可與移動式測繪平台100通訊，其中，合適的通訊協定例如為TCP/IP協定、DLC協定、IPX/SPX協定、HTTP協定等。資料後製系統200可對於初始資料122進行後製的資料處理，以便產出3D點雲地圖208。圖2繪示本發明一實施例在資料後製系統200中所配置的軟體架構之簡易方塊圖，圖3至圖5則繪示本發明一實施例在資料後製系統200中所執行之3D點雲地圖的產生方法流程圖。參閱圖2-5，資料後製系統200中可配置為包括一資料處理模組202及一點雲重建模組204的軟體架構。資料後製系統200中可執行步驟300，其可包括如圖4顯示的步驟302~314，將兩筆3D點雲接合產生經校正的3D點雲206，資料後製系統200接著可執行步驟400，其可包括如圖5顯示的步驟402~414，根據經校正的3D點雲206利用光達定位原理校正測繪平台原始定位誤差，如：修正原始點雲參考之GPS定位資料的誤差，後續資料後製系統200並可執行步驟500以疊合多筆經修正的3D點雲而建構含精確大地座標之3D點雲地圖。以下敘述將進一步詳述各步驟的內容。The data post-production system 200 is, for example, a personal computer, an industrial computer, or the like. The data post-production system 200 can obtain the initial data 122 from the data processing unit 110 and receive the input control point position 124 and the check point position 126. The initial data 122 includes, but is not limited to, the data obtained by the optical radar 104 including many 3D point cloud with multiple sampling points. The data post-production system 200 can obtain the initial data 122 through any suitable method. For example, the data post-production system 200 can communicate with the mobile surveying and mapping platform 100, where suitable communication protocols are, for example, TCP/IP, DLC, IPX/ SPX agreement, HTTP agreement, etc. The data post-production system 200 can perform post-production data processing on the initial data 122 to produce a 3D point cloud map 208. FIG. 2 shows a simple block diagram of the software architecture configured in the data post-production system 200 according to an embodiment of the present invention, and FIGS. 3 to 5 show the 3D implementation in the data post-production system 200 according to an embodiment of the present invention. The flow chart of the point cloud map generation method. 2-5, the data post-production system 200 can be configured as a software architecture including a data processing module 202 and a cloud reconstruction module 204. Step 300 can be performed in the data post-production system 200, which can include steps 302 to 314 as shown in FIG. , Which may include steps 402 to 414 shown in Figure 5, using the LiDAR positioning principle to correct the original positioning error of the surveying platform based on the corrected 3D point cloud 206, such as: correcting the error of the GPS positioning data referenced by the original point cloud, and subsequent data The post-production system 200 can perform step 500 to construct a 3D point cloud map with accurate geodetic coordinates by superimposing multiple corrected 3D point clouds. The following description will further detail the content of each step.

參閱圖4，初始步驟302可包括取得複數個3D點雲內至少一特徵點並給定各特徵點一精準量測的控制點之三維大地座標。舉例來說，可取得對應特徵的控制點與欲校正的兩筆3D點雲中對應相同特徵的取樣點，取樣點包括其3D位置資訊與時間資訊，係來自於移動式測繪平台100中的光學雷達104。特徵可為環境中對應此些3D點雲欲接合處的物件、圖樣或其他外觀特徵，如標線角落、電箱角落等，在本實施例中，可經由人工實地量測環境中的多個特徵取得其絕對大地座標作為控制點的位置124輸入至資料後製系統200，並可依需求選擇環境中適合的特徵作為控制點，舉例為道路上的特定圖案的特定位置、固定設置於環境中的特定物體的特定位置等，在此無限制。各個控制點之間的距離並無限制，可依據需求或實際特徵位置設定。圖6繪示一示例之3D點雲中的取樣點與控制點之位置關係圖。取樣點例如為圖6中所示的黑點P1、P2，控制點例如包括圖6中所示的白點CT1、CT2。在本實施例中，資料處理模組202可依據取樣點P1、P2的一標記資訊確認其分別與控制點CT1、CT2對應相同特徵，因此在3D點雲中找出第一組對應相同特徵的控制點CT1和取樣點P1以及第二組對應相同特徵的控制點CT2和取樣點P2。Referring to FIG. 4, the initial step 302 may include obtaining at least one feature point in a plurality of 3D point clouds and assigning each feature point a three-dimensional geodetic coordinate of a precisely measured control point. For example, the control points corresponding to the feature and the sampling points corresponding to the same feature in the two 3D point clouds to be calibrated can be obtained. The sampling points include their 3D position information and time information, which are derived from the optical system in the mobile surveying and mapping platform 100. Radar 104. The features can be objects, patterns, or other appearance features in the environment where these 3D point clouds are to be joined, such as corners of markings, corners of electrical boxes, etc. In this embodiment, multiple measurements in the environment can be performed manually. The feature obtains its absolute geodetic coordinates as the position of the control point 124 and inputs it to the data post-production system 200, and can select suitable features in the environment as the control point according to requirements, for example, a specific position of a specific pattern on a road, fixed in the environment The specific location of the specific object, etc., there is no restriction here. The distance between each control point is not limited and can be set according to requirements or actual feature positions. Fig. 6 shows an example of the positional relationship between the sampling points and the control points in the 3D point cloud. The sampling points are, for example, the black points P1 and P2 shown in FIG. 6, and the control points include, for example, the white points CT1 and CT2 shown in FIG. 6. In this embodiment, the data processing module 202 can confirm that the sampling points P1 and P2 correspond to the same features as the control points CT1 and CT2 respectively according to the marking information of the sampling points. The control point CT1 and the sampling point P1 and the second group of control points CT2 and the sampling point P2 corresponding to the same characteristics.

接著，在步驟304中，資料處理模組202可判斷距離各個控制點CT1、CT2一預定範圍（以圓圈標示）內要校正的控制點，如此可確認落在此預定範圍中的控制點為何以便進行下一步驟的校正。此預定範圍大小與形狀並無限制，可依據需求設定，舉例來說，預定範圍可為平面5公尺內高度差在10公尺內。Then, in step 304, the data processing module 202 can determine the control point to be corrected within a predetermined range (marked by a circle) from each control point CT1, CT2, so that it can be determined which control point falls within the predetermined range so as to Proceed to the next step of calibration. The size and shape of the predetermined range are not limited, and can be set according to requirements. For example, the predetermined range can be within 5 meters of a plane with a height difference of 10 meters.

接著，在步驟306中，資料處理模組202可依照對應特徵的控制點與取樣點的相對位置進行校正，更新取樣點的位置資訊。詳細地說，依照對應特徵點給定的控制點之三維大地座標之間的相對位置，對所有特徵點之間的取樣點位置做比例的水平、垂直移動校正。在此舉例資料處理模組202可依據控制點CT1和取樣點P1的相對位置計算得到一位置變換矩陣，並以此位置變換矩陣計算取樣點P1與預定範圍中待校正的取樣點的新的位置。位置變換矩陣對每一取樣點P1與預定範圍中待校正的取樣點施以權重係數，使得取樣點P1的新的位置與控制點CT1重合，而預定範圍中待校正的取樣點的位置改變幅度係對應各該取樣點與取樣點P1的距離。換言之，對於原先距取樣點P1越遠的取樣點，其新的位置越接近其原本的位置，並維持位於預定範圍以外的取樣點的位置不變，如此可達到較佳的平滑效果。在本實施例中，位置變換矩陣示例可為[dx,dy,dz]’，然無限制於此。接著，資料處理模組202可一組接續一組地繼續校正下一組的取樣點P2與預定範圍中待校正的取樣點，直至完成所有組的校正。圖7繪示經校正的3D點雲中的取樣點與檢核點之位置關係圖，其中的黑點即為經校正之取樣點。Then, in step 306, the data processing module 202 can perform calibration according to the relative positions of the control points of the corresponding features and the sampling points, and update the position information of the sampling points. In detail, according to the relative positions of the three-dimensional geodetic coordinates of the control points given by the corresponding characteristic points, the proportional horizontal and vertical movement corrections are made to the positions of the sampling points between all the characteristic points. In this example, the data processing module 202 can calculate a position transformation matrix based on the relative positions of the control point CT1 and the sampling point P1, and use the position transformation matrix to calculate the new position of the sampling point P1 and the sampling point to be corrected in the predetermined range. . The position transformation matrix applies a weight coefficient to each sampling point P1 and the sampling point to be corrected in the predetermined range, so that the new position of the sampling point P1 coincides with the control point CT1, and the position of the sampling point to be corrected in the predetermined range changes by the magnitude It corresponds to the distance between each sampling point and sampling point P1. In other words, for a sampling point that is farther away from the sampling point P1, its new position is closer to its original position, and the position of the sampling point outside the predetermined range remains unchanged, so that a better smoothing effect can be achieved. In this embodiment, an example of the position transformation matrix may be [dx,dy,dz]', but it is not limited thereto. Then, the data processing module 202 can continue to calibrate the sampling points P2 of the next group and the sampling points to be calibrated in the predetermined range one group after another, until the calibration of all groups is completed. FIG. 7 illustrates the positional relationship between the sampling points and the check points in the corrected 3D point cloud, in which the black dots are the corrected sampling points.

隨後，可進行步驟308，將欲接合的此些3D點雲在前述接合處接合，由於接合處已透過對應環境中相同特徵的控制點CT1、CT2和取樣點P1、P2進行校正，欲接合的3D點雲可在此獲得較佳的一致性。Then, step 308 can be performed to join the 3D point clouds to be joined at the aforementioned joints. Since the joints have been calibrated through the control points CT1, CT2 and sampling points P1, P2 of the same feature in the corresponding environment, the joints to be joined The 3D point cloud can get better consistency here.

接著，可進行步驟310，依據圖7顯示的例子，判斷經校正的3D點雲是否符合精度要求。在此可透過對應較佳位於控制點CT1、CT2連線正中間位置的特徵的一檢核點CP與經校正的3D點雲中對應相同特徵的取樣點P3的位置差是否落於一精度要求範圍內而得知是否符合精度要求。與控制點類似地，特徵可為環境中對應選定位置的物件、圖樣或其他外觀特徵。在本實施例中，可經由人工實地量測環境中的多個特徵取得其絕對大地座標作為檢核點CP的位置126輸入至資料後製系統200，並可依需求選擇環境中適合的特徵作為檢核點CP，舉例為道路上的特定圖案的特定位置、固定設置於環境中的特定物體的特定位置等，在此無限制。精度要求範圍可依據需求設定，如：平面誤差20cm內高度誤差30公分內。若位置差未超出此精度要求範圍時，即符合精度要求，進行步驟314，資料後製系統200產出經校正的3D點雲。若位置差超出此精度要求範圍時，即不符合精度要求，進行步驟312，取得更多的控制點及檢核點並重複進行步驟302-310。Then, step 310 can be performed to determine whether the corrected 3D point cloud meets the accuracy requirements according to the example shown in FIG. 7. Here, it can be determined whether the position difference between a check point CP corresponding to a feature preferably located in the middle of the connection between the control points CT1 and CT2 and a sampling point P3 corresponding to the same feature in the corrected 3D point cloud falls within an accuracy requirement. Within the range and know whether it meets the accuracy requirements. Similar to control points, features can be objects, patterns, or other appearance features corresponding to selected locations in the environment. In this embodiment, the absolute geodetic coordinates can be obtained by manually measuring multiple features in the environment as the location of the check point CP 126 and input to the data post-production system 200, and suitable features in the environment can be selected as required. The check point CP is, for example, a specific position of a specific pattern on a road, a specific position of a specific object fixedly installed in the environment, etc., and there is no limitation here. The accuracy requirement range can be set according to the requirements, such as: plane error within 20cm, height error within 30cm. If the position difference does not exceed the accuracy requirement range, the accuracy requirement is met, and step 314 is performed, and the data post-processing system 200 generates a corrected 3D point cloud. If the position difference exceeds the accuracy requirement range, that is, the accuracy requirement is not met, go to step 312 to obtain more control points and check points, and repeat steps 302-310.

接著，參閱圖5，進行步驟402，資料後製系統200可取得前述經校正的3D點雲與GPS定位資料。GPS定位資料可透過移動式測繪平台100的導航定位裝置102在GPS誤差較大的區域的一固定地點取得，較佳地，此固定地點可接近一特徵。特徵可為環境中的物件、圖樣或其他外觀特徵，舉例為道路上的特定圖案的特定位置、固定設置於環境中的特定物體的特定位置等，在此無限制。Next, referring to FIG. 5, step 402 is performed, and the data post-processing system 200 can obtain the aforementioned corrected 3D point cloud and GPS positioning data. The GPS positioning data can be obtained through the navigation and positioning device 102 of the mobile surveying and mapping platform 100 at a fixed location in an area with a large GPS error. Preferably, the fixed location can be close to a feature. The features can be objects, patterns, or other appearance features in the environment, such as a specific position of a specific pattern on a road, a specific position of a specific object fixed in the environment, etc., which are not limited here.

接著，進行步驟404，資料後製系統200可透過移動式測繪平台100的光學雷達104依據經校正的3D點雲對該特徵取得一光達定位結果，藉此修正來自定位系統的位置誤差。一般而言，光學雷達104可具備自行定位的功能，亦即光學雷達104可透過掃描和計算的方式判斷光學雷達104相對於特定物體或特徵之位置。故，步驟404中之光達定位結果為透過光學雷達104對該特徵掃描並使用一計算方式產生，為獨立於定位系統的資料估算光學雷達104所在的位置，其中，該計算方式例如包括迭代最近點（Iterative Closest Point，簡稱ICP）、迭代最近投影點（Iterative Closest Projected Point，簡稱ICPP）、正態分佈變換（Normal Distribution Transform，簡稱NDT）等。由於點雲資料已被控制點修正至大地座標上，光達定位的優勢即為有精準的距離量測特性參考有精準大地座標的點雲地圖從而推算當下含有大地座標的位置，此法在都市遮蔽嚴重的場景下可以補足衛星定位之不足，因此取得更精準的定位結果軌跡重建點雲地圖為本發明關鍵地思考方向。Then, in step 404, the data post-production system 200 can obtain an optical positioning result for the feature through the optical radar 104 of the mobile surveying and mapping platform 100 according to the corrected 3D point cloud, thereby correcting the position error from the positioning system. Generally speaking, the optical radar 104 can have a self-positioning function, that is, the optical radar 104 can determine the position of the optical radar 104 relative to a specific object or feature through scanning and calculation. Therefore, the LiDAR positioning result in step 404 is generated by scanning the feature through the optical radar 104 and using a calculation method to estimate the position of the optical radar 104 for data independent of the positioning system. The calculation method includes, for example, iterative nearest Point (Iterative Closest Point, ICP), Iterative Closest Projected Point (ICPP), Normal Distribution Transform (NDT), etc. Since the point cloud data has been corrected to the geodetic coordinates by the control points, the advantage of LiDAR positioning is that it has accurate distance measurement characteristics. Refer to the point cloud map with accurate geodetic coordinates to calculate the current position containing the geodetic coordinates. This method is used in urban areas. Under severely obscured scenes, the lack of satellite positioning can be supplemented. Therefore, obtaining a more accurate positioning result and reconstructing a point cloud map is the key thinking direction of the present invention.

接著，進行步驟406，資料後製系統200可比較光達定位結果與GPS定位資料的誤差，並接續執行步驟408，判斷光達定位結果與GPS定位資料的誤差是否超過一預定誤差範圍。若誤差未超出預定誤差範圍，接著進行步驟410，不修正GPS定位資料，如：不修正定位軌跡，並後續進行步驟414，產出重新建構之修正後精確的GPS定位軌跡。在實際的狀況中衛星定位地行走軌跡為時間序列包含定位結果經度緯度跟高度，而利用校正的點雲地圖及光達定位結果亦可在對應的時間序列上產出對應的經度緯度高度資訊。若在步驟408中的誤差超出預定誤差範圍，接著進行步驟412，依據該光達定位結果修正GPS定位資料，如：修正GPS定位軌跡，較佳地可提升精度至10公分之內。隨後，進行步驟414，依據修正後的定位資料產出重新建構定位軌跡，較佳地重新建構精確的定位軌跡。Then, in step 406, the data post-production system 200 can compare the error of the LiDAR positioning result with the GPS positioning data, and continue to execute step 408 to determine whether the error of the LiDAR positioning result and the GPS positioning data exceeds a predetermined error range. If the error does not exceed the predetermined error range, then proceed to step 410 without correcting the GPS positioning data, for example, not correcting the positioning trajectory, and then proceed to step 414 to generate a reconstructed and corrected accurate GPS positioning trajectory. In the actual situation, the walking track of the satellite positioning is a time series including the positioning result longitude, latitude and height, and the corrected point cloud map and the LiDAR positioning result can also be used to generate corresponding longitude, latitude and height information on the corresponding time series. If the error in step 408 exceeds the predetermined error range, then step 412 is performed to correct the GPS positioning data according to the LiDAR positioning result, such as correcting the GPS positioning track, preferably the accuracy can be improved to within 10 cm. Then, step 414 is performed to reconstruct the positioning track based on the corrected positioning data output, and preferably to reconstruct an accurate positioning track.

透過前述步驟302~314、402~414產生一個經校正的3D點雲並據此產生光達定位結果，再依光達定位結果與GPS定位資料的誤差修正定位軌跡之後，點雲地圖產生系統10可藉著移動式測繪平台100產生多個高精度的3D點雲。接著，在步驟500中，點雲重建模組204將每一高精度的3D點雲疊合以產生一高精度座標的3D點雲地圖208。Through the aforementioned steps 302~314, 402~414, a corrected 3D point cloud is generated and the LiDAR positioning result is generated accordingly. After the positioning track is corrected according to the error between the LiDAR positioning result and the GPS positioning data, the point cloud map generation system 10 The mobile surveying and mapping platform 100 can be used to generate multiple high-precision 3D point clouds. Next, in step 500, the point cloud reconstruction module 204 superimposes each high-precision 3D point cloud to generate a high-precision coordinate 3D point cloud map 208.

本發明建構3D點雲地圖的產生方法與系統，可透過控制點校正取樣點而產生經校正的3D點雲，並依據經校正的3D點雲修正定位資料的誤差，從而提升三維位置上的精確度，並將3D點雲疊合而產生具有高擬合、高精度位置的3D點雲地圖。The method and system for generating a 3D point cloud map constructed by the present invention can generate a corrected 3D point cloud by correcting sampling points through control points, and correcting errors in positioning data according to the corrected 3D point cloud, thereby improving the accuracy of the three-dimensional position The 3D point cloud is superimposed to produce a 3D point cloud map with high fitting and high-precision location.

本發明已經通過上述實施例進行了說明，但應當理解的是，上述實施例只是用於舉例和說明的目的，而非意在將本發明限制於所描述的實施例範圍內。此外本領域技術人員可以理解的是，本發明並不局限於上述實施例，根據本發明的教導還可以做出更多種的變型和修改，這些變型和修改均落在本發明所要求保護的範圍以內。本發明的保護範圍由附屬的申請專利範圍及其等效範圍所界定。The present invention has been described through the above-mentioned embodiments, but it should be understood that the above-mentioned embodiments are only for the purpose of example and description, and are not intended to limit the present invention to the scope of the described embodiments. In addition, those skilled in the art can understand that the present invention is not limited to the above-mentioned embodiments, and more variations and modifications can be made according to the teachings of the present invention, and these variations and modifications fall under the protection of the present invention. Within the range. The protection scope of the present invention is defined by the attached patent application scope and its equivalent scope.

10:3D點雲地圖產生系統 100:移動式測繪平台 102:導航定位裝置 104:光學雷達 110:資料處理單元 112:時間同步模組 114:資料紀錄模組 116:作業系統 122:初始資料 124:控制點的位置 126:檢核點的位置 200:資料後製系統 202:資料處理模組 204:點雲重建模組 206:3D點雲地圖 P1, P2, P3:取樣點 300-314, 400-414, 500:步驟10: 3D point cloud map generation system 100: Mobile surveying and mapping platform 102: Navigation and positioning device 104: Optical radar 110: data processing unit 112: Time synchronization module 114: data record module 116: operating system 122: initial data 124: Location of control points 126: Location of Check Point 200: data post-production system 202: Data Processing Module 204: Point cloud reconstruction module 206: 3D point cloud map P1, P2, P3: sampling point 300-314, 400-414, 500: steps

圖1繪示本發明一實施例所提供之含大地座標之3D點雲地圖產生系統之方塊圖。FIG. 1 is a block diagram of a 3D point cloud map generating system including geodetic coordinates provided by an embodiment of the present invention.

圖2繪示本發明一實施例在含大地座標之3D點雲地圖產生系統的資料後製系統中所配置的軟體架構之簡易方塊圖。FIG. 2 shows a simple block diagram of the software architecture configured in the data post-production system of the 3D point cloud map generating system containing geodetic coordinates according to an embodiment of the present invention.

圖3至圖5繪示本發明一實施例在資料後製系統200中所執行之含大地座標之3D點雲地圖的產生方法的流程圖。3 to 5 show a flowchart of a method for generating a 3D point cloud map containing geodetic coordinates executed in the data post-production system 200 according to an embodiment of the present invention.

圖6繪示3D點雲中的取樣點與控制點之位置關係圖。Fig. 6 is a diagram showing the positional relationship between sampling points and control points in the 3D point cloud.

圖7繪示本發明一實施例經重建的之含大地座標之3D點雲的取樣點與檢核點之位置關係圖。FIG. 7 is a diagram showing the positional relationship between sampling points and check points of a reconstructed 3D point cloud containing geodetic coordinates according to an embodiment of the present invention.

300,400,500:步驟300, 400, 500: steps

Claims

一種3D點雲地圖的產生方法，應用於包括多個取樣點的多筆3D點雲，包括：產生一經校正的3D點雲；根據該經校正的3D點雲，修正由一定位系統提供的定位資料之誤差；及產生一含大地座標之3D點雲地圖。A method for generating a 3D point cloud map is applied to a multi-stroke 3D point cloud including multiple sampling points, including: Generate a corrected 3D point cloud; According to the corrected 3D point cloud, correct the errors in the positioning data provided by a positioning system; and Generate a 3D point cloud map with geodetic coordinates.

如請求項1所述的方法，其中產生一經校正的3D點雲的步驟包括：對複數個3D點雲分別進行下列步驟：取得該複數個3D點雲內至少一特徵點並給定各該特徵點一精準量測的控制點之三維大地座標；及依照對應該特徵點給定的該控制點之該三維大地座標之間的相對位置，對所有該特徵點之間的取樣點位置做比例的水平、垂直移動校正。The method according to claim 1, wherein the step of generating a corrected 3D point cloud includes: Perform the following steps on multiple 3D point clouds: Obtain at least one feature point in the plurality of 3D point clouds and assign the three-dimensional geodetic coordinates of each of the feature points to a precisely measured control point; and According to the relative position between the three-dimensional geodetic coordinates of the control point given corresponding to the characteristic point, a proportional horizontal and vertical movement correction is made to the position of the sampling point between all the characteristic points.

如請求項2所述的方法，其中產生一經校正的3D點雲的步驟包括：判斷對應該控制點的一預定範圍內要校正的複數個取樣點；及判斷經校正的該複數個3D點雲的該複數個取樣點是否符合精度要求。The method according to claim 2, wherein the step of generating a corrected 3D point cloud includes: Determine the plurality of sampling points to be corrected within a predetermined range corresponding to the control points; and It is determined whether the plurality of sampling points of the plurality of corrected 3D point clouds meet the accuracy requirements.

如請求項1所述的方法，其中根據該經校正的3D點雲，修正由一定位系統提供的定位資料之誤差之步驟包括：依據該經校正的3D點雲取得一光達定位結果；比較該光達定位結果與該定位資料的誤差；及修正該定位資料。The method according to claim 1, wherein the step of correcting errors in positioning data provided by a positioning system according to the corrected 3D point cloud includes: Obtain a LiDAR positioning result according to the corrected 3D point cloud; Compare the error of the LiDAR positioning result with the positioning data; and Correct the positioning data.

如請求項4所述的方法，其中根據該經校正的3D點雲，修正由一定位系統提供的定位資料之誤差之步驟包括：判斷超過一預定誤差範圍時，修正該定位資料，且判斷不超過該預定誤差範圍時，不修正該定位資料；及產出重新建構之該修正後精確定位軌跡。The method according to claim 4, wherein the step of correcting errors in positioning data provided by a positioning system according to the corrected 3D point cloud includes: When it is judged that it exceeds a predetermined error range, the positioning data is corrected, and when it is judged that it does not exceed the predetermined error range, the positioning data is not corrected; and The reconstructed accurate positioning trajectory is produced after the correction.

如請求項4所述的方法，其中該光達定位結果係由一光學雷達的迭代最近點、迭代最近投影點、正態分佈變換之一計算方式所產生。The method according to claim 4, wherein the LiDAR positioning result is generated by one of an optical radar's iterative closest point, iterative closest projection point, and normal distribution transformation.

一種3D點雲地圖產生系統，包括：一移動式測繪平台，其中裝設有一導航定位裝置、一光學雷達及一資料處理單元，其中該導航定位裝置接收來自一定位系統提供的定位資料，該資料處理單元可接收來自導航定位裝置的該定位資料及來自該光學雷達的多筆3D點雲，每一筆3D點雲包括多個取樣點；及一資料後製系統，接收該複數個3D點雲及複數個控制點的位置，該資料後製系統係被配置以執行以下步驟：產生一經校正的3D點雲；根據該經校正的3D點雲，修正該定位資料之誤差；及產生一含大地座標之3D點雲地圖。A 3D point cloud map generation system, including: A mobile surveying and mapping platform, which is equipped with a navigation and positioning device, an optical radar and a data processing unit, wherein the navigation and positioning device receives positioning data provided by a positioning system, and the data processing unit can receive the navigation and positioning device Positioning data and multiple 3D point clouds from the optical radar, each 3D point cloud including multiple sampling points; and A data post-production system receives the positions of the plurality of 3D point clouds and a plurality of control points. The data post-production system is configured to perform the following steps: Generate a corrected 3D point cloud; According to the corrected 3D point cloud, correct the error of the positioning data; and Generate a 3D point cloud map with geodetic coordinates.

如請求項7所述的3D點雲地圖產生系統，其中該資料後製系統係被配置以執行產生一經校正的3D點雲的該步驟更包括：對複數個3D點雲分別進行下列步驟：取得該複數個3D點雲內至少一特徵點並給定各該特徵點一精準量測的控制點之三維大地座標；及依照對應該特徵點給定的該控制點之該三維大地座標之間的相對位置，對所有該特徵點之間的取樣點位置做比例的水平、垂直移動校正。判斷經校正的該複數個3D點雲的該複數個取樣點是否符合精度要求；及產出一經校正的3D點雲。The 3D point cloud map generation system according to claim 7, wherein the data post-production system is configured to perform the step of generating a corrected 3D point cloud further comprising: Perform the following steps on multiple 3D point clouds: Obtain at least one feature point in the plurality of 3D point clouds and assign the three-dimensional geodetic coordinates of each of the feature points to a precisely measured control point; and According to the relative position between the three-dimensional geodetic coordinates of the control point given corresponding to the characteristic point, a proportional horizontal and vertical movement correction is made to the position of the sampling point between all the characteristic points. Determine whether the plurality of sampling points of the plurality of corrected 3D point clouds meet the accuracy requirements; and Produce a corrected 3D point cloud.

如請求項7所述的3D點雲地圖產生系統，其中該資料後製系統係被配置以執行根據該經校正的3D點雲，修正該定位資料之誤差的該步驟更包括：依據該經校正的3D點雲取得一光達定位結果；比較該光達定位結果與該定位資料的誤差；判斷超過一預定誤差範圍時，修正該定位資料，且判斷不超過該預定誤差範圍時，不修正該定位資料；及產出重新建構之該修正後精確定位軌跡。The 3D point cloud map generation system according to claim 7, wherein the data post-production system is configured to execute the step of correcting the error of the positioning data according to the corrected 3D point cloud further including: Obtain a LiDAR positioning result according to the corrected 3D point cloud; Compare the error of the LiDAR positioning result with the positioning data; When it is judged that it exceeds a predetermined error range, the positioning data is corrected, and when it is judged that it does not exceed the predetermined error range, the positioning data is not corrected; and The reconstructed accurate positioning trajectory is produced after the correction.

如請求項7所述的3D點雲地圖產生系統，其中該導航定位裝置包括一全球定位系統接收器及一慣性量測單元，該全球定位系統接收器接收來自衛星之全球定位系統定位資料，該慣性量測單元連續估算即時位置。The 3D point cloud map generation system according to claim 7, wherein the navigation and positioning device includes a global positioning system receiver and an inertial measurement unit, the global positioning system receiver receives global positioning system positioning data from satellites, the The inertial measurement unit continuously estimates the real-time position.