1307860 九、發明說明： 【發明所屬之技術領域】 本發明涉及一種電腦輔助檢測系統及方法，尤其是關 於一種點雲格式解析系統及方法。 【先前技術】 爲了提高和保證産品品質，或是研究市場上出現的新 産品以輔助生産製造流程，企業常常需要對一些産品進行 檢測。 ® 早期對於受測物件的檢驗，一般採用特定的檢驗儀 器，經人工比對後確定檢測值。近年來，隨著電腦硬體性 能的提高及價格的降低，電腦在受測物件檢驗活動中被大 量的引入，因而提高了檢驗的速度和準確性。其做法一般 是使用掃描機台掃描受測物件，獲得由多個三維離散點組 成的點的集合，一般稱之爲點雲，將點雲資料匯入電腦， 執行相應軟體對點雲資料進行處理，比如量測點雲，色階 Φ 比對等，從而實現對受測物件之檢驗。 然而目前的點雲檢測系統所能識別的點雲格式具有 一定局限性，例如大多數檢測系統不支援Metris公司的 MBF(Metris basic file)點雲槽案格式，而Metris公司在高 端掃描系統市場佔有率較高。且沒有一個完整的點雲模型 對點雲物件進行描述。 【發明内容】 鑒於以上内容，有必要提供一種點雲格式解析系統及 方法，對不同格式的點雲檔案進行解析，建立完整的點雲 7 1307860 物件模型，以方便點雲檢測。 料广!=解析系蘇，該系統包括應用飼服器及. ==，點_，所述應用词服= 括.點雲檔案接收模組㈣服盎包 斷模組，用於判斷點_ “貝4庫接收點雲檔案；判 :據點雲檔案的格式解析點雲檔案以搜索二:】於 雲物件模型建立媪鈿 m’6為貝成’及點 資訊儲存至點雲物件類點雲物件類、將所述點雲 類所存儲的點雲資訊建立_=2 ^據點雲物件 其中，所述解析模組包 置解析參數，包括在不_ ，數攻置子^組，用於設 時的步長及將二進位資二=的點雲檀案中搜索點雲資訊 換長度；資料格式轉換tr爲用戶可識別的資料時的轉 稽案以二驗格式RU，用於將非文本格式的點雲 度轉換成用戶i域二進位資料按設定的轉換長 於按設定的步長在點=料’及點雲資訊搜索子模級，用 ^ 、 _’、关檐案中搜索點雲資訊。 模組，用於===物件模型建立模組包括：建立子 儲的點雲資訊捷*件類’及用於根據點雲物件類所存 於初始化點雲物件件模型；陣列初始化子模組，用 陣列名稱、維Π存儲陣列，包括初始化存儲陣列的 解析所得的點L貝料類型；及資訊添加子模組，用於將 一種點兩B貝訊儲存至點雲物件類的相應存儲陣列。 從資料庫接彳=方法，該方法包括以下步驟:(A) ‘ π檔案，（B)判斷該點雲檔案是否爲文本 1307860 Γ二二Γ 文本格式，則解析該文本格式 ^雲檔案不爲文本格式，則以二進位格 案，並解析得到的二進位格式的點雲檔 利用本發明，可對不同抵 取掃描物件的點雲資訊，建立樓案進行解析’獲 點雲進行檢測。點雲物件模型，方便用戶對 【實施方式】 的硬=示，！本發明點雲格式解析系統較佳_ 的硬體、構圖。該祕包括點雲掃描υ卜應用 網路3、應用伺服器4及資料庫 、、 件得到物件的點雲資料，資料庫5:=裝置1掃描物 的點雲資料健存爲相應的點雲檀荦，庫^描裝置1得到 個功能模組及CAD賴，對㈣4包括多 互動建立點雲的物件模型’應用終端2提供 能:t:c:用戶通過網路3調用應用伺服器4上的功 7模組和广軟體，㈣3可以爲企 網路 ㈤軸t)、網際網路（Internet)或其他類型網路。 的功=τ，係本發明點雲格式解析系統較佳實施例 判斷ϋ。應用舰器4包括：點雲檔案接收模組1〇、 16、二：：式解析模組14、點雲物件模型建立模組 處理权組18及輸入模組20。 播案點雲檔案接收模組10用於接收來自資料庫$的點雲 9 1307860 判斷模組12用於判斷點雲接收模組1〇 標案的格式，點雲槽案的格式可分爲文本牧所接收的點雲 式兩大類。其中應用最f遍的文本袼 1和非文本格 ASCH ( American standard c〇de ^槽案之一爲1307860 IX. Description of the Invention: [Technical Field] The present invention relates to a computer-aided detection system and method, and more particularly to a point cloud format analysis system and method. [Prior Art] In order to improve and guarantee product quality, or to research new products on the market to assist manufacturing processes, companies often need to test some products. ® In the early inspection of the tested object, a specific inspection instrument is generally used, and the detection value is determined by manual comparison. In recent years, with the improvement of computer hardware performance and price reduction, computers have been introduced in large quantities in the inspection activities of tested objects, thus improving the speed and accuracy of inspection. The method generally uses a scanning machine to scan the object to be tested, and obtains a set of points consisting of a plurality of three-dimensional discrete points, generally called a point cloud, and the point cloud data is sent to the computer, and the corresponding software is processed to process the point cloud data. For example, the measurement point cloud, the gradation Φ comparison, etc., to achieve the inspection of the object under test. However, the point cloud format that the current point cloud detection system can recognize has certain limitations. For example, most detection systems do not support Metris' MBF (Metris basic file) point cloud slot format, and Metris has a high-end scanning system market. The rate is higher. And there is no complete point cloud model to describe point cloud objects. SUMMARY OF THE INVENTION In view of the above, it is necessary to provide a point cloud format analysis system and method for parsing point cloud files of different formats, and establishing a complete point cloud 7 1307860 object model to facilitate point cloud detection. Wide range of materials! = Analytical system, the system includes the application of the feeding device and . ==, point _, the application word service = include. Point cloud file receiving module (four) service ang package module, used to judge points _ "Bei 4 library receives point cloud file; judge: According to the format of the point cloud file, the point cloud file is searched for two:] The cloud object model is built 媪钿m'6 for Beicheng' and the point information is stored to the point cloud object type point cloud. The object class, the point cloud information stored in the point cloud class is established _=2 ^ the point cloud object, wherein the parsing module includes parsing parameters, including in the _, number attack subgroup, for setting The step size of the time and the point cloud information in the point cloud 2 = search point cloud information exchange length; the data format conversion tr is the user-recognizable data when the audit file is in the second format RU, for non-text The format of the point cloud is converted into the user i domain binary data according to the set conversion longer than the set step size in the point = material 'and point cloud information search sub-module level, use ^, _', search for the point cloud Information. Module, for === object model building module includes: creating sub-storage point cloud information According to the point cloud object class stored in the initialization point cloud object model; the array initialization sub-module, using the array name, the dimension storage array, including the initialization of the storage array analysis point L material type; and information adding sub-modules, The utility model is used for storing a point B and B information to a corresponding storage array of the point cloud object class. From the database connection method, the method comprises the following steps: (A) 'π file, (B) determining whether the point cloud file is For the text 1307078 Γ二二Γ text format, the text format is parsed. If the cloud file is not in the text format, the binary file is parsed and the binary file format of the binary format is parsed. Scan the point cloud information of the object, and establish a building case for analysis. The point cloud is used for detection. The point cloud object model is convenient for the user to perform the hard implementation of the [implementation method], and the hardware of the point cloud format analysis system of the present invention is better. The composition includes the point cloud scanning application network 3, the application server 4 and the database, and the point cloud data of the object, and the database 5:= the point cloud data of the device 1 is saved as Corresponding point cloud sandalwood, library scanning device 1 gets a function module and CAD Lai, (4) 4 includes multiple interactive object cloud model object application 'application terminal 2 provides energy: t:c: user calls application through network 3 The power 7 module and the wide software body on the server 4, (4) 3 can be the network (5) axis t), the Internet (Internet) or other types of networks. The work = τ, is the point cloud format analysis system of the present invention. The preferred embodiment determines that the application vessel 4 includes: a point cloud file receiving module 1〇, 16, 2: a type analysis module 14, a point cloud object model establishing module processing right group 18, and an input module 20. The cloud file receiving module 10 is configured to receive a point cloud 9 1307860 from the database $. The determining module 12 is configured to determine the format of the point cloud receiving module 1 〇 standard file. The format of the point cloud slot file can be divided into text grazing. The two types of point cloud types received. One of the most frequently used texts 袼 1 and non-text ASCH (American standard c〇de ^ one of the slots is
Interchange,美國資訊交換標準碼），非文=^formation 案以在高端掃描彡統市場佔有轉高的^的點雲稽 典型代表。 F點雲檔案爲Interchange, the US information exchange standard code), the non-text=^formation case is a typical representative of the point cloud in the high-end scanning system. F point cloud file is
14料㈣料觀 f進行解析，獲取點雲資訊，所述點雲資訊包^對點雲檔 屬性及各點座標值。s & I括點雲名稱、 舌件 值其中，點雲屬性包括點雲Μ 冉 4 所處圖層 '總點數及點雲最大雲 ^式解析触14包括：參數設 =界點等。 式轉換子模組U2及 且141、資料格 子模組⑷用於設置解搜索143。參•置 檔案中搜索時的步長。括“置在不同格式點雲 資料時的轉換長度等爲用〜二 文本袼式的點雲標子模紙142用於將非 轉換成用戶可識別的資料式開啓，並將二Μ資料 根據設置參數在點雲檔案中索子模級143用於 點雲物件模型建立模电16索^負訊。 、 ^建立點騎件賴型，用於根據射⑽點 創建點雲物件類，及用建立子模級峋 化點雲物件類的在㈣車歹J初始化子模組161，·〜資 件類的存儲陣列，包括陣列名稱、維數用，初始 資料類 型，·及資訊添加子模組162 錯存至相應存錯陣列。 '將解析所得的點雲吹 . Θ貝訊 處理模組18用於對所建 步處理’包括去雜點、調整點：：點雲物件模型壤行進〜 或消除多個點雲物件重曼的部^件的顏色、亮度，合併 系統物件㈣輪人至 生成對應的點雲圖形。 主CAt) 、，如圖3所示，係本發明點带故j14 material (four) material view f to analyze, to obtain point cloud information, the point cloud information package ^ point cloud file attributes and coordinates of each point. s & I include the cloud name and tongue value. Among them, the point cloud attribute includes the point cloud Μ 所 4 where the layer is located. 'The total number of points and the maximum cloud of the point cloud. The type of resolution touch 14 includes: parameter setting = boundary point. The conversion sub-modules U2 and 141 and the data grid sub-module (4) are used to set the solution search 143. Steps to set the step size in the file search. Including "the conversion length when setting the point cloud data in different formats, etc., the point cloud standard paper 142 using the ~2 text type is used to open the non-converted data format that can be recognized by the user, and the data is set according to the setting. In the point cloud file, the parameter 143 is used for the point cloud object model to establish the mode circuit 16 cable ^ negative signal. , ^ Establish the point riding type, used to create the point cloud object class according to the shot (10) point, and use to establish The sub-module leveling point cloud object class is in the (four) rut J initialization sub-module 161, the storage class of the component class, including the array name, the dimension number, the initial data type, and the information adding sub-module 162. Misplaced to the corresponding misplaced array. 'The point cloud will be parsed. The ΘBai processing module 18 is used to process the built steps' including the de-noise, adjustment point:: point cloud object model soil travel ~ or eliminate The color and brightness of the parts of the plurality of point cloud objects are combined with the system object (4) to generate the corresponding point cloud pattern. The main CAt), as shown in Fig. 3, is the point of the present invention.
首先，點雲产索垃 π α式解析方法的埯程R (步顿咖=接收資料庫5儲存的點4衆 式，如1 = (步驟st 格㈣#難，搜索賴取點雲資^ 雲資點雲物件模型建立模心根據搜索到= 所建立^ 物件的模型（步驟Sl6)。處理模版18對 除多個Γ雲物件模型進行處理，包括去雜點、合併气消 將處I;，件重疊的部分等(步驟叫輸入模= 的點+谈的點雲物件模型輸人至CAD系統⑽，生成對麻 幾計^形(步驟S20)。所述CAD系統是一種電腦輔:、 據需要t如如〇°^、灿/£、31)如等。用戶也可根 使用，^解析得到的點雲資訊儲存到新的檔案中便於後續 如儲存到由用戶命名的一個WORD標案中。 只 袼式在3:12,若判斷模組12判斷該點雲播案不爲文本 雲橋案Γ 轉換子模組二2以二進位格式打開該點 （V驟S24)。格式解析模組14解析得到的二進位 11 1307860 式點雲槽案（步驟S26)。隨後流程轉入S16。 建立t圖4所示，係圖3中步驟S26的細化流程圖。首先，First of all, the point cloud R solute π α-analysis method of the process R (steps of the coffee = receive the database 5 stored point 4 public, such as 1 = (step st grid (four) # difficult, search for the point cloud ^ The cloud asset point cloud object model establishes a model according to the searched = model of the object created (step S16). The processing template 18 processes the plurality of cloud object models, including denoising, combining gas elimination, and I; , the overlapping parts of the parts, etc. (the step is called input mode = point + talk point cloud object model is input to the CAD system (10), and the pair is generated (step S20). The CAD system is a computer auxiliary: According to the needs of t such as 〇 ° ^, Can / £, 31), etc. Users can also use the root, ^ parsed point cloud information stored in a new file for subsequent storage, such as a WORD standard named by the user In the case of only 3:12, if the judgment module 12 judges that the point cloud broadcast case is not a text cloud bridge case, the conversion sub-module 2 2 opens the point in the binary format (V step S24). The group 14 parses the obtained binary 11 1307860 type point cloud slot case (step S26). Then the flow goes to S16. The establishment t is shown in Fig. 4, which is shown in Fig. 3. The detailed flowchart of step S26. First,
ClaSS)(^^ 陳列+彳化子彳歧161初始化點雲物件類的存儲 件名稱t陣列名稱、維數、料類型，例如：將點雲物 么％的存儲陣列命名爲“SetName” 、設置維數爲〜 f、毁置資料類型爲字元（eha〇;將點雲物件中點的存= '列命名爲“AddPoim” 、設置維數爲三維、設置資料麵 =爲浮點型（float)等（步驟sl〇2)。參數設置子模組又幻 λ置將二進位資料轉換爲用戶可識別的資料（如字元、十 進位數字字）時的轉換長度，如以電腦中可單獨處理的最 小單位一個位元組爲單位轉換長度，二進位資料的最小單 位爲比特（bit)，lbyte=8bits，及設置在轉換後的資料中嫂 索時的步長，如一個位元組（步驟sl〇3)。資料格式轉換 子模組142將檔案中的二進位（binary)資料按預設的轉 換長度轉換爲用戶可識別的資料（步驟Si〇4)。點雲資訊 搜索子模組143依預設的步長讀取轉換後點雲檔案中的資 料（步驟S105)。判斷模組12判斷是否搜索到點雲名稱資 訊，如表示點雲名稱的“ New scan(l)，’ 、 M New scan(2)’’ 、“Camera 1”等字串（步驟S106)。若搜索到 表示點雲名稱的資訊，則資訊添加子模組162儲存搜索到 的點雲名稱至點雲物件類中物件名稱的存儲陣列，如將點 雲名稱“ New scan(l) ”儲存到物件類的存儲陣列 “SetName”中（步驟S107)。判斷模組12判斷是否搜索 12 1307860 到所述點雲的屬性資訊’如表示 等的字串“Layer” 、 “以―:雲的圖層、顏色、亮度 咖）。若搜索到表示所述點=居“Li，等（步驟 子模組162儲存搜索到的點，屬性資訊’則資訊添加 應屬性的儲存陣列（步驟S1:/生資訊至點雲物件類中相 繼續搜索，_關斷是:資訊搜索子模組⑷ 資訊，如點雲中點的_數，^索到所述點雲的點座標 束點的子串如“start point”、 /、、。 _)。若搜索到點座標資訊，_ =二^(步驟 從檔案中總點數 ：-貝减索子模組143 aw 4疋榡誌起始點的字串後，分 二人項取點雲的所有點的χγζ座榡值，如每次讀取⑽依 (lfloat=4bytes=32bits)型資料，一 〇at 料值依讀取順序H個三 餘值，絲各點χγζ__ 2 儲陣列，如“AddPoint” （步驟_千類干點的存 是否到達_尾(步驟===， 槽案中還有未讀取的資料，則返回步驟·，點雲^ 索子模版143繼續搜索其他點雲的資訊。 、成拽 在步驟Sl〇6，若判斷模纟且u判斷未搜索到點 資訊’則直接進入步驟S108。 稱 在步驟S108，若判斷模組12判斷未搜索到點雲 資訊’則直接進入步驟S110。 *生 在步驟S110，若判斷模組12判斷未搜索到點座樟次 汛，即點雲資訊搜索子模組143未搜索到點座標資訊，身 則 13 1307860 轉入步驟S112。 • 如圖5所示，係圖3中步驟S14的細化流程圖。步驟 S201、S202分別類似步驟sl〇1、sl〇2。參數設置子模組 141設置在文字檔案中搜索的步長，例如若該文本格式的 點雲檔案爲ASCII點雲檔案，則可設置步長爲一行，即以 相鄰兩個換行符或回車鍵之間的距離作爲搜索時的步長 (步驟S203)。點雲資訊搜索子模組ι43在檔案中依設定 的步長讀取檔案中的資料（步驟S204)。判斷模組12判斷 是否搜索到點雲名稱資訊，類似於圖4中步驟sl〇6 (步驟 S205)。步驟S206、S207分別類似於步驟S107、S110。若 在步驟S205判斷模組判斷搜索到點座標資訊，則點雲資 訊搜索子模組143依設定的搜索步長依次讀取點雲的所有 點的XYZ座標值，如步長爲一行，則每次讀取一行資料， 資訊添加子模組162將每行中以一個特殊字串，如“*，，、 等’或η (n>=1)個空格所夾的三個數值依讀取順 _ 序赋給一個三維離散點作爲XYZ座標值，並將各點χγΖ 座標值儲存到點雲物件類中點的存儲陣列（步驟S208)。 判斷模組12判斷是否到達檔案結尾（步驟S209)。若未到 達播案結尾，即檔案中還有未讀取的資料，則返回步驟 S204’點雲資訊搜索子模組ι43繼續搜索其他點雲的資訊。 在步驟S205 ’若判斷模組12判斷未搜索到點雲名稱 資訊，則進入步驟S207。 在步騍S207，若判斷模組u判斷未搜索到點座標資 訊’即點雲資訊搜索子模組143未搜索到點座標資訊，則 1307860 進入步驟S209。 關於點雲的資訊，點雲檔案中不一定有每一塊點雲的 名稱、屬性資訊，但只要有點雲，一定有點座標資訊。若 在搜索過程中，點雲資訊搜索子模組143只搜索到點雲的 點座標資訊，則在步驟S16，點雲物件模型建立模組16爲 該點雲賦予默認的名稱和屬性。 如圖6所示，以對MBF點雲檔案的解析爲例說明本 發明點雲格式解析系統及方法。點雲檔案接收模組1 〇從資 — 料庫5接收非文本格式的點雲檔案後，將該點雲檔案臨時 儲存在應用伺服器的一個記憶體空間（步驟S401)。資料 格式轉換子模組142以二進位格式打開該點雲檔案後，以 一個位元組（lbyte=8bits)爲單位長度將二進位資料轉換 爲用戶可識別的資料，如字元、十進位數字字（步驟 S402)。點雲資訊搜索子模組143以一個位元組爲步長在 轉換後的點雲檔案中搜索點雲檔案格式資訊（步驟 φ S403 )。判斷模組12判斷點雲檔案中是否有Metris資訊（步 驟 S404)。若找到如“Metris”、“Metris_file_format” 等 字串，則判斷模組12判斷該點雲檔案爲MBF點雲檔案， 點雲資訊搜索子模組143繼續搜索該MBF檔案中所包含 的點雲資訊：包括點雲名稱、屬性及座標資訊，“New scan” 、“ Camera”等表示點雲名稱常用的特定字串；搜 索“Layer” 、“Color” 、“Light”等表示點雲圖層、顏 色、亮度等屬性常用的特定字串；及搜索各點雲中所有點 的XYZ座標值（步驟S405)。點雲物件模型建立模組16 15 1307860 根據搜索到的點雲資訊建立該MBF點雲檔案中所包含的 點雲物件的模型（步驟S406 )。 在步驟S404，若判斷模組12判斷未搜索到Metris資 訊，則結束流程。 本發明雖以較佳實施例揭露如上，然其並非用以限定 本發明。任何熟悉此項技藝者，在不脫離本發明之精神和 範圍内，當可做更動與潤飾，因此本發明之保護範圍當視 後附之申請專利範圍所界定者為準。 •【圖式簡單說明】 圖1係發明點雲格式解析系統較佳實施例的硬體架構 圖。 圖2係本發明點雲格式解析系統較佳實施例的功能模 組圖。 圖3係本發明點雲格式解析方法的流程圖。 圖4係圖3中步驟解析得到的二進位格式的點雲檔案 φ 的細化流程圖。 圖5係圖3中步驟解析該文本格式的點雲檔案的細化 流程圖。 圖6以對MBF點雲檔案的解析爲例說明本發明點雲 格式解析系統及方法。 【主要元件符號說明】 點雲掃描裝置 應用終端 網路 1 2 3 16 1307860 應用伺服器 4 資料庫 5 點雲檔案接收模组 10 判斷模組 12 格式解析模組 14 參數設置子模組 141 資料格式轉換子模組 142 點雲資訊搜索子模組 143 點雲物件模型建立模組 16 陣列初始化子模組 161 資訊添加子模組 162 建立子模組 163 處理模組 18 輸入模組 20 CAD系統 40ClaSS)(^^ Display + 彳 彳 161 161 Initialization point cloud object class storage name t array name, dimension, material type, for example: the point cloud object% of the storage array named "SetName", set Dimension is ~ f, the data type is destroyed as character (eha〇; save the point cloud of the point cloud object = 'column named "AddPoim", set the dimension to 3D, set the data surface = float type (float ), etc. (step sl1). The parameter setting sub-module and the illusion are converted to the length of the user-recognizable data (such as characters, decimal digits), such as in the computer can be separate The minimum unit of processing is one unit of conversion length, the minimum unit of binary data is bit (bit), lbyte=8bits, and the step size when setting the data in the converted data, such as a byte ( Steps s1〇3) The data format conversion sub-module 142 converts the binary data in the file into user-recognizable data according to a preset conversion length (step Si〇4). The point cloud information search sub-module 143 reads the converted point cloud file in the preset step size Data (step S105). The judging module 12 judges whether or not the point cloud name information is searched, such as "New scan(l), ', M New scan(2)'', "Camera 1", etc. indicating the name of the point cloud. (Step S106). If the information indicating the name of the point cloud is searched, the information adding sub-module 162 stores the searched point cloud name to the storage array of the object name in the point cloud object class, such as the name cloud name "New scan( l) "stored in the storage array "SetName" of the object class (step S107). The judging module 12 judges whether to search 12 1307860 to the point cloud attribute information 'such as the string "Layer", " : cloud layer, color, brightness coffee.) If the search indicates that the point = "Li, etc. (step sub-module 162 stores the searched point, attribute information), the information is added to the storage array of the attribute (step S1) : / Health information to the point cloud object class continues to search, _ shutdown is: information search sub-module (4) information, such as the number of points in the point cloud, ^ to the point cloud point point of the point cloud Strings such as "start point", /,, . _). If you search for a point coordinate News, _ = two ^ (steps from the total number of points in the file: - 贝 索 子 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 143 Value, such as each reading (10) according to (lfloat = 4bytes = 32bits) type data, one 〇 at material value according to the reading order H three residual values, silk points χ ζ ζ __ 2 storage array, such as "AddPoint" (step _ thousand Whether the existence of the class dry point reaches the _ tail (step ===, there is still unread data in the slot case, then return to the step ·, point cloud ^ scorpion template 143 continue to search for information about other point clouds. In step S1, if the mode is judged and u judges that the point information is not found, the process proceeds directly to step S108. If it is determined in step S108 that the determination module 12 determines that the point cloud information has not been searched, the process proceeds directly to step S110. * In step S110, if the judging module 12 judges that the point coordinates are not searched, the point cloud information search sub-module 143 does not search for the point coordinate information, and the body 13 1307860 proceeds to step S112. • As shown in FIG. 5, it is a refinement flowchart of step S14 in FIG. Steps S201 and S202 are similar to steps sl1 and sl2, respectively. The parameter setting sub-module 141 sets the step size searched in the text file. For example, if the point cloud file of the text format is an ASCII point cloud file, the step size may be set to one line, that is, two adjacent line feeds or a carriage return. The distance between the keys is taken as the step size at the time of the search (step S203). The point cloud information search sub-module ι43 reads the data in the file in the file according to the set step size (step S204). The judging module 12 judges whether or not the point cloud name information is searched, similar to the step sl6 in Fig. 4 (step S205). Steps S206 and S207 are similar to steps S107 and S110, respectively. If it is determined in step S205 that the module determines to search for the point coordinate information, the point cloud information search sub-module 143 sequentially reads the XYZ coordinate values of all points of the point cloud according to the set search step length, for example, if the step size is one line, then Reading one line of data at a time, the information adding sub-module 162 reads the three values in a special string, such as "*,,, etc." or η (n>=1) spaces, in each line. The sequence is assigned to a three-dimensional discrete point as an XYZ coordinate value, and each point χγΖ coordinate value is stored in the storage array of the midpoint of the point cloud object class (step S208). The determination module 12 determines whether the file end is reached (step S209). If the end of the broadcast is not reached, that is, there is still unread data in the file, then return to step S204' point cloud information search sub-module ι43 to continue searching for information of other point clouds. In step S205, if the judgment module 12 determines not If the point cloud name information is searched, the process proceeds to step S207. In step S207, if the module u judges that the point coordinate information is not searched, that is, the point cloud information search sub-module 143 does not search for the point coordinate information, the step 130778 enters the step. S209. About point cloud Information, point cloud files do not necessarily have the name and attribute information of each point cloud, but as long as there is a little cloud, there must be some coordinate information. If during the search process, the point cloud information search sub-module 143 only searches for the point cloud point. For the coordinate information, in step S16, the point cloud object model building module 16 assigns a default name and attribute to the point cloud. As shown in FIG. 6, the parsing of the MBF point cloud file is taken as an example to illustrate the point cloud format parsing of the present invention. System and method: The point cloud file receiving module 1 receives the point cloud file in the non-text format from the asset-repository 5, and temporarily stores the point cloud file in a memory space of the application server (step S401). After the format conversion sub-module 142 opens the point cloud file in a binary format, the binary data is converted into user-recogable data, such as characters and decimal digits, in units of a byte (lbyte=8 bits). (Step S402) The point cloud information search sub-module 143 searches for the point cloud file format information in the converted point cloud file in steps of one byte (step φ S403 ). The determining module 12 determines the point cloud file. Whether there is Metris information in the case (step S404). If a string such as "Metris" or "Metris_file_format" is found, the judging module 12 judges that the point cloud file is an MBF point cloud file, and the point cloud information search sub-module 143 continues. Search for the point cloud information contained in the MBF file: including point cloud name, attribute and coordinate information, "New scan", "Camera", etc., which are used to indicate the specific string of the point cloud name; search for "Layer", "Color", "Light" or the like indicates a specific character string commonly used for attributes such as a point cloud layer, color, brightness, and the like; and XYZ coordinate values of all points in each point cloud are searched (step S405). The point cloud object model building module 16 15 1307860 creates a model of the point cloud object included in the MBF point cloud file based on the searched point cloud information (step S406). In step S404, if the judging module 12 judges that the Metris information is not searched, the flow ends. The present invention has been described above by way of a preferred embodiment, and is not intended to limit the invention. The scope of the present invention is defined by the scope of the appended claims, unless otherwise claimed. • [Simplified Schematic Description] Fig. 1 is a hardware architecture diagram of a preferred embodiment of the inventive point cloud format analysis system. Figure 2 is a functional block diagram of a preferred embodiment of the point cloud format analysis system of the present invention. 3 is a flow chart of a method for analyzing a point cloud format of the present invention. FIG. 4 is a detailed flowchart of the point cloud file φ in the binary format obtained by the step analysis in FIG. Figure 5 is a detailed flow chart of the step cloud file in Figure 3 for parsing the text format. FIG. 6 illustrates a point cloud format analysis system and method of the present invention by taking an analysis of an MBF point cloud file as an example. [Main component symbol description] Point cloud scanning device application terminal network 1 2 3 16 1307860 Application server 4 database 5 point cloud file receiving module 10 judgment module 12 format analysis module 14 parameter setting sub-module 141 data format Conversion sub-module 142 point cloud information search sub-module 143 point cloud object model building module 16 array initialization sub-module 161 information adding sub-module 162 building sub-module 163 processing module 18 input module 20 CAD system 40
1717