1327296 九、發明說明 月3日修正替換頁 【發明所屬之技術領域】 本發月係關於-種提升電子收費系統之扣款資料與執法 資料自動匹配作業成功率的方法，尤其是指在vps電子收費 系統中’運用GPS相對漂移向量校正、車輛時間差執法點預 L、可令錯式車牌比對等三種方法以準確預估裝載〇此的 車輛通過VPS電子收費系統執法線所行經的車道(如:内側、 中側、外側、路料車道），進而與在執法線上裝載在各車道 的攝影機所取像的車牌影像辨識結果做匹配，以解決一般 VPS電子收費系統匹配準確度不易提升的問題。 【先前技術】 -般VPS電子㈣系統運作原理為：VPS電子收費車載 電滕（On Board computer，〇BC)搭载Gps衛星定位裝置與 咖S無線通訊裝置’該0BC會自動透過Gps定位裝置讀取 現在車輛所在的位置，t車輛行車朗事先定義好的收費區 域時’OBC便會自動以GPRS無線通訊的方式通報後端帳務 系統’後端帳務系統將根據〇BC的識別碼與使用者事先登 錄的車型種類、扣款帳戶等資料，進行扣除通行費用之程 序’·當車輛經過收費區域的同時’區域内的執法線（搭載照相 機與感應器的銷架橫向線段稱之為執法線）配置有影像執法 系統’該影像執法系統會對所有通過收費區域的車輛予以取 像’並且自動做車牌影像辨識。倘若該車輛未搭裁〇BC或 5 1-327296 件邮日修正替換頁 未開啟OBC通過收費區域時，則後端僅收到其執法影像資 料，沒有收到扣款資料，即會將該車判定為未缴費之車辆， 需要透過監理系統取得該車之登記住址以進行後續的通行 費追缴事宜。 6亥VPS電子收費系統中’其交易扣款系統與影像執法系 統係分別獨立運作，其分別產生的扣款資料與執法資料將會 分別的傳送到後端帳務系統，再以批次整合匹配的方式將扣 款資料與執法資料匹配起來，以確認通過收費區域的車輛何 者有正確扣放、何者未裝置車載機，必須由執法影像追繳通 行費用。此匹配的過程受限於GPS衛星定位訊號的漂移、車 輛在收費區的行駛路線與執法取像車牌辨識率等種種因素 的衫響，導致後端帳務系統匹配的準確度不甚理想需要額 外藉助人工將無法匹配的資料重新校正、匹配，藉以提高匹 配準確度，以降低誤罰用路人的機率。 為解決定位訊號的漂移問題，一般認為有二種解決方 式：利用差分定位訊號（DGPS)校正漂移或是加裝特定短距離 通訊裝置（如：微波或紅外線）之車機與車道系統。然而，這 兩種方式都有缺點，前者是收費區必須很大，而通訊費用很 两；後者則需要兩套長短距離通訊裝置，其建置之成本又太 南〇 由上述VPS電子收費的流程描述得知，後端的整合匹配 rjz/zy〇 fk光日修 系統在VPS電子收參系紅士 ------- 1 电于收費系統中佔有關鍵 的結果準確度不佳，則大晋㈣ 右整合匹配 人工… 置的執法資料與扣款資料需要藉由 人工輔助系統匹配的工作， ^ # ^ 的成本遠高於系統處理 所侍到手續費用，則& VPS系 買万式將變得窒礙難 订路^甚者’若錯誤的匹配結果導致誤罰正常通行扣款的 被追繳通行f，則收費正確性遭受到質疑，可能導致 VPS電子收費失敗收場。 目前用來匹配執法影像車 平砰辨識與扣政資料的依據包 含：通過扣款區域時OBC直始 機所回傳的車輛所在GPS位置、 OBC代碼、回傳時間·輿摄 興攝&機所在位置、執法f彡像取像時 間與執法影像車牌辨識後的車 交耵早牌旎碼。根據這些資訊匹配出 來的結果，可能發生下述問題： 漂移與OBC回傳時間和後 置和時間為主要依據的匹 1. k料匹配錯誤：GP S的位置 端時間的差距都有可能導致以位 配程式產生錯誤的結果。 2·資料無m車牌轉㈣OCR車牌㈣之準確度 很難達到百分之百’導致於執法影像資料無法與扣款資料匹 配成功，必須仰賴後續的人q配處理機制才能解決問題。 為了解決扣款系統與執法系統之匹配準確度問題，先前 的技術包3以差分定位訊號校正與加裝特定短距離通訊裝 置兩種方法’其方法與優缺點簡述如下： 7 1327296 月〉日修正替換頁 1.將差分定位訊號（DGPS)利用GPRS通訊送到〇BC， OBC使用此訊號校正GPS定位漂移，使OBC之定位精準度 可以達到5米以内。此方法優點是〇bc不需要加裝任何設 備’僅需要軟體的升級；缺點則要達到足夠高的定位精準度 必須要長時間的傳送DGPS訊息，如此收費區域可能需要長 達數公里，並且通訊費用將會大幅提高。 2·在VPS系統上加裝短距離通訊（DSRC)裝置如紅外線或 微波短程通訊，以便車道系統可以精準的定位車輛行經收費 區之軌跡。此種解決方式的優點是可以解決Gps定位漂移的 問題； 缺點則是車上單元配置長短距離兩套通訊裝置成本太 車道系統也需要較高的成本。1327296 IX. Description of the invention on the 3rd of the month. Replacement page [Technical field of invention] This is a method for automatically matching the deduction data and law enforcement data of the electronic toll collection system, especially in vps In the charging system, three methods, such as GPS relative drift vector correction, vehicle time difference law enforcement point pre-L, and misplaced license plate comparison, are used to accurately estimate the lanes through which the vehicles carrying the VPS electronic toll collection system operate. : the inner side, the middle side, the outer side, the roadway lane), and then match the license plate image recognition result taken by the camera loaded in each lane on the law enforcement line to solve the problem that the matching accuracy of the general VPS electronic toll collection system is not easy to be improved. [Prior Art] - The general VPS electronic (four) system operating principle is: VPS electronic toll car on board (On Board computer, 〇 BC) equipped with GPS satellite positioning device and coffee S wireless communication device 'The 0BC will automatically read through the GPS positioning device Now the location of the vehicle, when the vehicle is driving a predefined charging area, 'OBC will automatically notify the back-end accounting system by means of GPRS wireless communication'. The back-end accounting system will be based on the identification code of the user and the user. The type of the vehicle to be registered in advance, the debit account, etc., and the procedure for deducting the toll fee '·When the vehicle passes through the toll area, the law enforcement line in the area (the horizontal section of the pin holder with camera and sensor is called the law enforcement line) It is equipped with an image law enforcement system. The image law enforcement system will take pictures of all vehicles passing through the toll area and automatically perform license plate image recognition. If the vehicle is not ruling BC or 5 1-327296, the postal date correction replacement page does not open the OBC through the charging area, then the back end only receives the law enforcement image data, and if the deduction information is not received, the vehicle will be For vehicles that are judged to be unpaid, it is necessary to obtain the registered address of the vehicle through the supervision system for follow-up toll recovery. In the 6 Hai VPS electronic toll collection system, its transaction deduction system and image law enforcement system operate separately, and the debit data and law enforcement data generated separately will be transmitted to the back-end accounting system separately, and then batch-matched and matched. The method is to match the debit information with the law enforcement data to confirm that the vehicle passing through the toll area has the correct deduction and the vehicle is not installed. The toll fee must be recovered by the law enforcement image. The process of matching is limited by the drift of GPS satellite positioning signals, the driving route of the vehicle in the toll area and the recognition rate of the law enforcement image license plate, etc., which leads to the accuracy of the matching of the back-end accounting system is not ideal. Manually recalibrate and match unmatched data to improve matching accuracy to reduce the chance of falsely penalizing passers-by. In order to solve the drift problem of the positioning signal, it is generally considered that there are two solutions: the differential positioning signal (DGPS) is used to correct the drift or the vehicle and the lane system of a specific short-distance communication device (such as microwave or infrared) are added. However, both of these methods have disadvantages. The former is that the toll area must be large, and the communication cost is very high. The latter requires two sets of long and short distance communication devices, and the cost of its construction is too small. The description shows that the integration of the back end matches the rjz/zy〇fk light day repair system in the VPS electronic receiving department Reds------- 1 electricity in the charging system occupies key results, the accuracy is not good, then Dajin (4) The right integration matches the manual... The law enforcement data and the debit data need to be matched by the manual assistance system. The cost of ^ # ^ is much higher than the processing fee of the system processing, then the & VPS system will change. It is difficult to make a reservation. If the wrong match results in a false penalty, the correctness of the normal charge will be questioned, and the correctness of the charge will be questioned, which may result in the failure of the VPS electronic charge. At present, the basis for matching the ID card identification and deduction information of the law enforcement image includes: the GPS position of the vehicle returned by the OBC straight-through machine when the deduction area is used, the OBC code, the return time, the camera, and the camera The location, law enforcement, and the image of the law enforcement image license plate after the car exchanged the early card weight. According to the results of the matching of these information, the following problems may occur: drift and OBC return time and post- and time-based are the main basis of 1. k material matching error: GP S position end time difference may lead to The bit program produces incorrect results. 2. The information does not change the license plate (4) The accuracy of the OCR license plate (4) It is difficult to reach 100%. As a result, the law enforcement image data cannot be matched with the debit data successfully. It must rely on the follow-up person's processing mechanism to solve the problem. In order to solve the problem of matching accuracy between the debit system and the law enforcement system, the previous technology package 3 uses differential positioning signal correction and installation of specific short-range communication devices. The methods, advantages and disadvantages are briefly described as follows: 7 1327296月〉日Correction Replacement Page 1. The differential positioning signal (DGPS) is sent to 〇BC using GPRS communication. OBC uses this signal to correct the GPS positioning drift, so that the positioning accuracy of OBC can reach less than 5 meters. The advantage of this method is that 〇bc does not need to install any equipment's only software upgrades are required; the disadvantage is that to achieve high enough positioning accuracy, it is necessary to transmit DGPS messages for a long time, so the charging area may take several kilometers and communication The cost will increase dramatically. 2. Install a short-range communication (DSRC) device such as infrared or microwave short-range communication on the VPS system, so that the lane system can accurately locate the trajectory of the vehicle through the toll zone. The advantage of this solution is that it can solve the problem of GPS positioning drift; the disadvantage is that the distance between the two units of the vehicle is long and the cost of the two sets of communication devices is too high.
設計design
種運用GPS相對漂移向量校GPS relative drift vector
以準確預估裝載OBC的車輛通過 比對等三種方法 VPS電子收費系統執法線 8 所仃經的車道(如··内側、中側、外側、路肩等車道)，進而 與在執法線上裝載在各車道的攝影機所取像的車牌影像辨 為結果做匹配，以解決—般vps電子收費系統匹配準確度不 易提升的問題’俾而達到正確判斷及扣款之目的。 本發明之次-目的係在於提供―種可提升扣款資料與執 法資枓自動匹配準確度，無需透過DGps校正或加裝短距離 通訊裝置’藉以提出一套成本低廉之電子收費系統之扣款資 料與執法資料自動匹配作業成功率的方法。 次可達成上述發明目的之電子收"' 統之扣款資料與執法 貝料自動匹配作業成功率的方法，係運用Gps相對漂移向量 校正、車輪時間差執法點預測與可容錯式車牌比對等三種方 法，以準綠預估裝載OBC的車輛通過vps電子收費系統執 法線所行經的車道（如··内側、中側、外側、路肩等車道）， 進而與在執法線上裝載在各車道的攝影機所取像的車牌影 像辨識結果做匹配，以解決—般vps電子收費系統匹配準確 度不易提升的問題》上述三種技術與功能簡述如下 KGPS相對漂移向量校正·將車機回報的㈣位置做一 相對你移ΐ校正’以求得較為精禮的車輛位置，以便推估車 翻行經執法線時所行敬的車道（如：内側、中側中外側、外 側、路肩等車道）。 2.車輛時間差執法點預測：Gps取樣速度為每秒一次， 9 作>月3曰修正替換頁 通過執法線時OBC不一定剛好讀取到⑽位置，透過内插 法的方式推估計算車㈣過執法線的實際座標位置與速 度，以彌補GPS取樣頻率不足而可能造成的位移誤差。 °、3式車牌比對.執法影像經過OCR車牌辨識的準 確率無法達到百分之百，可容供十击始、，/ J令錯式車牌以行經收費區域的車 輛所註冊之車牌_為基礎，tt對車牌號碼之辨識率，辨識 率達到一定標準值以上即算正確。 運用上述二種之方式，將可降低Gps漂移產生的誤差與 誤判車輛通過執法線所產生的資料自動匹配錯誤，並且透過 可容錯式車牌比對提高扣款資料與執法資料匹配比對的成 功率。 【實施方式】 請參閲圖一所示，為本發明接^ 1个赏3杈升電子收費系統之扣款資 料與執法資料自動匹配作業忐汾奎认士、Α 。奸糸成功率的方法之（VehicleIn order to accurately predict the OBC-loaded vehicles, the three lanes of the VPS electronic toll collection system law enforcement line 8 (such as the inside, the middle side, the outer side, the shoulder and other lanes), and then loaded on the law enforcement line The license plate image taken by the camera of the lane is identified as the result to match, so as to solve the problem that the matching accuracy of the vps electronic toll collection system is not easy to improve, and the purpose of correct judgment and deduction is achieved. The second objective of the present invention is to provide an automatic matching accuracy of the information of the deduction of the deduction and the law enforcement, without the need to correct or install the short-distance communication device through the DGps, so as to propose a deduction of a low-cost electronic toll collection system. The method of automatically matching the data and law enforcement data to the success rate of the job. The method of automatically matching the data of the electronic receipt and the law enforcement of the above-mentioned inventions is to use the Gps relative drift vector correction, the wheel time difference law enforcement point prediction and the fault-tolerant license plate comparison. There are three methods to predict the lanes (such as the inner side, the middle side, the outer side, the shoulders, etc.) of the vehicles carrying the OBC through the vps electronic toll system law enforcement line, and then the cameras loaded in the lanes on the law enforcement line. The image recognition result of the image taken by the image is matched to solve the problem that the matching accuracy of the vps electronic toll collection system is not easy to be improved. The above three techniques and functions are briefly described as follows: KGPS relative drift vector correction, and the position of the vehicle is returned as a (four) position. Relative to your shifting corrections, in order to obtain a more refined vehicle position, in order to estimate the lanes that the vehicle travels when passing through the law enforcement line (eg, inside, mid-side, outside, shoulder, shoulder, etc.). 2. Vehicle time difference law enforcement point prediction: Gps sampling speed is once per second, 9 for > month 3 曰 correction replacement page OBC does not necessarily just read the (10) position when passing the law enforcement line, push the estimated calculation method by interpolation method (4) The actual coordinate position and speed of the law enforcement line to compensate for the displacement error that may be caused by insufficient GPS sampling frequency. °, 3 type license plate comparison. The accuracy rate of law enforcement image through OCR license plate identification can not reach 100%, can be used for 10 strikes, / / J wrong license plate with the license plate registered by the vehicle in the toll area _ based, tt For the identification rate of the license plate number, the recognition rate is more than a certain standard value. Using the above two methods, the error caused by the Gps drift can be reduced and the data generated by the vehicle passing the law enforcement line can be automatically mismatched, and the success rate of the matching data and the law enforcement data can be improved through the fault-tolerant license plate comparison. . [Embodiment] Please refer to Figure 1 for the automatic matching of the deduction information and law enforcement data of the 1st reward and 3 liter electronic toll collection system of the present invention. The method of success rate of traitors (Vehicle
Positioning System，VPS)電子收費系統架構示意圖主要包 括搭載有GPS定位裝置2與GPRS/3G遠距離通訊裝置3的 VPS電子收費車上單元i(即為〇BC)、收費站區的執法系統 4(包含攝影機與車牌辨識模組5)、扣款模組6與帳務系統7 等所構成。 請參閱圖二所示，為該提升VPS電子收費系統之扣款資 料與執法資料自動匹配作業成功率的方法之GPS相對漂移 1327.296 愧工月知fc替換頁 向里校正示意圖’係利用裝置在攝影機上的GPS的漂移向 量’與在同一時間内通過執法線8車輛之vps電子收費車上 單元1取彳于的GPS位置漂移向量差距很小的特性。利用攝影 機已知的正確座標位置，對照攝影機上同一時間所傳回來的 GPS座標位置可以計算出漂移向量利用此漂移向量反過來 修正同一時間内車輛GpS所傳回的座標位置，即可得出車輛 當時的正確座標位置。 請參閱圖三所示，為該提升電子收費系統之扣款資料與 '、·貝料自動配作冑成功率的方法之車輛通過執法線相 對位置預測不意圖，由於車輛上之vps電子收費車上單元1 透過GPS取得座標的頻率為每秒―次在車速高達1嫌祕^ 上的间速公路用路環境，每秒車輛的位移高達π 78公 尺，VPS電子收鲁鱼, 費車上早70 1取樣GPS的地點可能與執法線 8相距超過甚遠’卩VPS電子收費車上單元！回報的㈣座 標為基準勢必料㈣料確性，因此本發明㈣通過執法 線8前後兩次的gps取描私尤，t ± 樣所仔到的車輛位置，行經方向與速 度’經過GPS相對漂移向量校正後取得正確的座標位置，再 利用内插法的方式計算車輛軌跡與執法線8的交叉點，此點 即是車輛通過執法線8時 与 攝〜機所取像的座標，用以比對 攝影機的座標位置，即可 寸算出車輛所行駛的車道（内侧、中 側、_外側、外侧、 月）。内插法的推估方法如下，假設車 1 ^/296 月3日修正替換頁 輕在行經跨越執法線8的兩回報點（時間差距一秒）的座標位 時1速度為⑷……’川與⑷从^冲内插法想要 求出該車輛通過執法線8的位置、時間、速度為 _鄭斗則⑽雜以藉由求執法線8與兩回報點所 形成的直線之交點敌理 Λ ^ 取侍，ν0與t0則可以藉由内插法計算牛 頓第二運動定律計算而得。 、取得車輛行經執法線8所行駛的隸位置與車道後，執 、'、統取像並做車牌號碼影像辨識，職結果與車機所預 先登錄的車牌號碼做進一步比對，以確認是否為裝置vps電 子收費車上單元1車機之車辆通過收費區域》由於車牌影像 〇cR辨識結果受限於光線、取像角度與車牌位置是否有遮蔽 物或污損等外在因切㈣料碼辨識之準確度，本發明提 出可容錯式車牌比對的方法，可以用來解決此問題。執法影 像經過OCR車牌辨識的準確率無法達到百分之百，無法2 識的車牌字元以特定符號代替（ex‘，，？，，）；可容錯式的車牌比對 藉由機率統計的方式降低車牌辨識錯誤無法與扣款資料作 匹配的比率進而提升匹配的成功率，如：車機所註冊的車 牌號碼為【綱234】之車輪經過執法系統4取像辨識車牌 號瑪之結果為【Α?·1234】’其準確度達到83 33%，同一時間 的類似車牌通過該執法線8之車輛機率微乎其微，故可視為 辨識正確。 12 1327296 在-般的VPS電子收費系統中，後端帳務系統7可以用 來將扣认資料與執法資料相匹配的資料並不多’包含州電 子收費車上單兀i車機所回傳的車輛所在GPS位置、则電 子收費車上單元！代碼、回傳時間、攝影機所在位置、執法 影像取像時間與執法影像車牌辨識後的車牌號碼，整合匹配 程式必須依據上述資料將已扣款的資料與執法影像匹配起 來，並將已扣款的車輛與重複取像㈣像資料過料，剩下 ㈣車牌影像就是未裝置VPS電子收費車上單元！車機（或 是裝置OBC故障或未開機）之車輛， ^ 神帳務處理人貝根據車牌 資料連線監理站資料庫取得用女 渾取仵用路人住址，以進行後續之通行 費催款動作。 是以’本發明係可運用GPS相對漂移向量校正車輛時 間差執法點㈣與可容錯式車料料三種技術方法中一 種或混合二種以上之技術方法來提升電子收費系統中扣款 資料與執法資料自動匹配作業之準確度„ 本發明所提供之提升電子收費系統之扣款資料與執法資 料自動匹配作業成功率的方法，與其他習用技術相互比較 時’更具有下列之優點： 般方式實做的自動匹配方法，因為Gps訊號漂移 與車牌辨識率料度並非百分之百，其自動匹配的準確度不 '，必須仰賴大量的人工處理校正；本發明所提出的三種技 13 1327296 ^•月》日修正替換頁 術方法可以大幅提昇準確度。 _ 2. 本發明所提出之解決方法比較以差分訊號校正〇ps位 置方式，可以節省大量的通信費用，並且不需要設定大範圍 的收費區域（差分訊號校正GPS的方法需要很大的收費站區 範圍’以便後端傳送足夠多的DGPS訊號供OBC修正GPS)。 3. 本發明所提出之解決方法比較以加裝短距離通訊裝置 方式，可以大幅降低OBC成本與〇BC耗電量，並且可以節 省車道系統的建置費用。 上列詳細說明係針對本發明之一可行實施例之具體說 明，惟該實施例並非用以限制本發明之專利範圍，凡未脫離 本發明技藝精神所為之等效實施或變更，均應包含於本案之 專利範圍中。 綜上所述’本案不但在技術思想上確屬創新，並能較習 用物品增進上述多項功效’應已充分符合新穎性及進步性之 法定發明專利要件，爰依法提出申請，懇請貴局核准本件 發明專利申請案，以勵發明，至感德便。 【圖式簡單說明】 請參閱以下有關本發明一較佳實施例之詳細說明及其附 圖，將可進一步瞭解本發明之技術内容及其目的功效；有關 該實施例之附圖為： 圖一為本發明提升電子收費系統之扣款資料與執法資料 1327296 自動匹配作業成功率的方法… 動 正示意圖，·以 圖二為該提升電子收7；費系統架構示意圖，· ㈣作業m 費系統之扣款資料與執法資料自 匹配作業成功率的方法之Gps相對漂移向量校 圖三為該提升電子收費系統之扣款資料與執法資料自動 匹配作業成功率的方法之車輛通過執法線相對位置預測示 意圖。 【主要元件符號說明】 1 VPS電子收費車上單元 2 GPS定位裝置 3 GPRS/3G遠距離通訊裝置 4執法系統 5車牌辨識模組 6扣款模組 7帳務系統 8執法線 9執法點 15Positioning System, VPS) The schematic diagram of the electronic toll collection system mainly includes the VPS electronic toll collection unit i (ie 〇BC) equipped with the GPS positioning device 2 and the GPRS/3G remote communication device 3, and the law enforcement system 4 of the toll station area ( The camera and license plate recognition module 5), the deduction module 6 and the accounting system 7 are included. Please refer to Figure 2 for the GPS relative drift of the method for automatically matching the deduction data and law enforcement data of the VPS electronic toll collection system. The relative drift of the method is shown in the camera. The drift vector of the GPS on the GPS transmission drift vector of the VPS electronic toll car on the law enforcement line 8 vehicle at the same time is very small. Using the correct coordinate position known to the camera, the GPS vector position transmitted at the same time on the camera can be used to calculate the drift vector. Using this drift vector to correct the coordinate position returned by the vehicle GpS at the same time, the vehicle can be obtained. The correct coordinate position at the time. Please refer to Figure 3 for the purpose of predicting the relative position of the vehicle through the law enforcement line for the deduction of the electronic toll collection system and the method of automatically selecting the success rate of the 'bean and the shell material, because the vps electronic toll vehicle on the vehicle The upper unit 1 obtains the coordinates of the coordinates through the GPS every second - the speed of the road is as high as 1 speed. The displacement of the vehicle is up to π 78 meters per second, and the VPS electronically receives the fish. The location of the early 70 1 sampling GPS may be more than the distance of the law enforcement line 8 '卩VPS electronic toll car on the unit! The (4) coordinates of the return are unavoidable. (4) The material is confirmed. Therefore, the present invention (4) takes the gps of the law enforcement line 8 twice before and after, and the position of the vehicle, the direction and speed of the passing through the GPS relative drift vector After correcting, obtain the correct coordinate position, and then use the interpolation method to calculate the intersection of the vehicle trajectory and the law enforcement line 8. This point is the coordinates of the image taken by the camera when the vehicle passes the law enforcement line 8 for comparison. The coordinates of the camera can be used to calculate the lane in which the vehicle is traveling (inside, in the middle, in the outside, in the outside, in the moon). The estimation method of the interpolation method is as follows. Assume that the vehicle 1 ^/296 March 3 corrects the replacement page light when the coordinate position of the two return points (time difference one second) crossing the law enforcement line 8 is 1 (4)... And (4) from the ^ impulse interpolation method to find the location, time, and speed of the vehicle passing the law enforcement line 8 _ Zheng Dou (10) mixed with the line of law formed by the law line 8 and the two points of the point of the enemy ^ Served, ν0 and t0 can be calculated by interpolation to calculate Newton's second law of motion. After obtaining the position and lane of the vehicle traveling through the law enforcement line 8, the operator will perform the image identification of the license plate number, and the job result will be further compared with the license plate number registered in advance by the vehicle to confirm whether it is The device vps electronic toll car on the unit 1 car machine through the toll area" because the license plate image 〇 cR identification results are limited by the light, the image angle and the license plate position whether there is a shield or stain, etc. (4) material code The accuracy of the identification, the present invention proposes a method of fault-tolerant license plate comparison, which can be used to solve this problem. The accuracy of law enforcement images through OCR license plate recognition cannot be 100%. The license plate characters that cannot be recognized are replaced by specific symbols (ex',,?,,); the fault-tolerant license plate comparison reduces the license plate identification by probability statistics. The error can not match the deduction data to increase the success rate of the match. For example, the license plate number registered by the car is the squad of the squad 234, and the result of the identification of the license plate number by the law enforcement system 4 is [Α?· 1234] 'The accuracy is 83 33%. At the same time, similar license plates pass through the law enforcement line 8 and the probability of the vehicle is very small, so it can be regarded as correct identification. 12 1327296 In the general VPS electronic toll collection system, the back-end accounting system 7 can be used to match the denomination data with the law enforcement data, and there is not much information included in the state electronic toll car. The GPS location of the vehicle, then the electronic toll car on the unit! The code, the return time, the location of the camera, the law enforcement image acquisition time and the license plate number after the law enforcement image license plate identification, the integration matching program must match the debited data with the law enforcement image based on the above information, and the debited The vehicle and the repeated image (4) image material are over, and the remaining (4) license plate image is the unit that is not equipped with the VPS electronic toll car! The vehicle (or the equipment OBC is faulty or not turned on), ^ God account processing person Bay according to the license plate data connection supervision station database to obtain the use of female pick-up passers-by address, for subsequent toll collection dunning action . The invention is based on the fact that the invention can use the GPS relative drift vector to correct the vehicle time difference law enforcement point (4) and the fault-tolerant vehicle material one of three technical methods or to mix two or more technical methods to improve the charge data and law enforcement data in the electronic toll collection system. Accuracy of automatic matching operation „ The method for improving the automatic matching operation rate of the electronic toll collection system and the law enforcement data provided by the invention has the following advantages when compared with other conventional technologies: Automatic matching method, because Gps signal drift and license plate recognition rate is not 100%, the accuracy of automatic matching is not ', must rely on a large number of manual processing correction; the three techniques proposed by the present invention 13 1327296 ^•月》 The page method can greatly improve the accuracy. _ 2. The solution proposed by the present invention compares the 〇ps position mode by differential signal, which can save a lot of communication costs, and does not need to set a wide range of charging areas (differential signal correction GPS) Method requires a large toll station area range for the back end to transmit There are enough DGPS signals for OBC to correct GPS. 3. The solution proposed by the present invention is compared with the short-distance communication device, which can greatly reduce the OBC cost and the BC BC power consumption, and can save the construction of the lane system. The detailed description above is a detailed description of one of the possible embodiments of the present invention, and is not intended to limit the scope of the present invention. It is included in the patent scope of this case. In summary, the 'this case is not only innovative in terms of technical thinking, but also can improve the above-mentioned multiple functions compared with the conventional articles'. It should fully comply with the statutory invention patent requirements of novelty and progressiveness. To apply, you are requested to approve the invention patent application, in order to invent the invention, to the point of feeling. [Simplified description of the drawings] Please refer to the following detailed description of a preferred embodiment of the present invention and its accompanying drawings, which may further The technical content of the present invention and the purpose of the present invention are as follows: Figure 1 is a diagram of the present invention for improving the electronic toll collection system. Section data and law enforcement data 1327296 method of automatically matching the success rate of the operation... Dynamic schematic diagram, Figure 2 shows the upgrade electronic receipt; fee system architecture diagram, (4) Self-matching of the deduction data and law enforcement data of the operation m fee system The Gps relative drift vector of the method of success rate is the schematic diagram of the relative position prediction of the vehicle through the law enforcement line for the method of automatically increasing the credit rate of the electronic toll collection system and the law enforcement data. [Main component symbol description] 1 VPS Electronic toll car on the unit 2 GPS positioning device 3 GPRS/3G remote communication device 4 law enforcement system 5 license plate recognition module 6 deduction module 7 accounting system 8 law enforcement line 9 law enforcement point 15