200901099 九、發明說明： 【發明所屬之技術領域】 本發月關於種動態物辨識系統，尤指一種以攝影機取得純影像 圖檔來判斷該物之身型或速料之辨識系統。 【先前技術】 在A路上對車輛作車型、車速及車牌_識祕多具有 像賴裝置及—雷達啦裝置賴成。圖像拍齡置至少具備 了-照像機及-閃光燈，在雷達猶裝置制到車輛超速時，會控制 圖像拍攝裝置啟動以取得超速車輛照片資料。 此一傳統辨識系統在使用時，由於電達測速裝置要發射一光束至 車輛上，例如雷射或紅外線等，然後才可測得車輛是否超速。這使得 車輛上可裝置一偵測機器事先通知駕駛，以躲避超速；這一現像使得 辨識系統防止超速的效果大打折扣。 有時’圖像拍攝裝置在拍攝時必需使用閃光燈，這突來的閃光有 f " 時會造成駕駛的驚嚇，反而造成交通事故。是以傳統的辨識系統實有 待改善。 由於MJPEG (Moving Pictures Experts Group)技術提供，它是 一種基於靜態圖像壓縮技術JPEG演變而成的動態圖像壓 縮技術。MJPEG能生成序列化的運動圖像，過程中不考慮 視訊流内不同晝面之間的變化，只單獨針對各晝面進行壓 縮，此種壓縮方式非常適合靜態畫面，解析度可從 352x240、704x480到1280x1024。因此，自攝影機取得的 5 200901099 連續多補圖檔，每-禎圖權所揭示影像都將是清晰可辨識 的。而本發明將利用此_技術以改善車輛辨識系統。 【發明内容】 本發明主要目的係在提供—種動態物辨識系統，利用—攝影機在 該動Μ勿仃經路從上设定一段攝影範圍在該動態物進入該攝影範圍 時4貞測裝置控綱攝影機拍攝連續影像圖樓，經由一衛星定位系 Γ. ，统⑽），取得每一圖標之經緯度數據及時間數據，以成為每-圖檔 之樓名。每-圖檔經由-套格裝置先將每—影像圖檔套上虛擬格線， 而該虛擬格線是以路徑上固定標記物長度或其他固定純度為參考而 置；^型比對裝置將攝f鐵取得之動態物之全身影像圖檔，利用 影像比對方式，出觸祕之身型。—速度計錢置將所拍攝的 至少兩禎以上圖檔相比對出距離差及時間差，以計算出該動態物之速 度。 本發明由於以純影像來判斷動態物之身型及速度，例如應用在火 I, 車、公路上行進車輛及航行船艦之判視上，只需在行經路徑上適當位 置架設攝影機即可，不會改變路徑上的環境及週邊設備。尤其，應用 在車輛時’不需要閃光燈及發出光束訊號’即可辨識車型及車速以 減少駕駛受驚而發生意外之機會，且使得車輛内無法偵測到該辨識系 統。 本發明依據此技術，也可將身型辨識及速度辨識分成獨立的兩個 獨立系統，來分別辨識身塑及速度。 以下，將依據圖面所示之實施例而詳加說明本發明系統特點及功 6 200901099 效。 【實施方式】 請參見第一圖，本發明所提供的動態物辨識系統s該動態物可以 疋火車、車輛或船艇等；本發明主要在動態物行經的路徑上適當位置 架設了至少—攝影機’該攝影機設定-段攝影範圍L，該攝影範圍L至 v可拍攝動祕之全身長度及—段行進長度，以取得連續多福高解晰 度的影像圖檔’例如以MJPEG技術拍攝。 一偵測裝置在偵測到動態物進入攝影範圍時控制攝影機開始拍 攝。 一衛星定位系統（GPS)用來取得每一拍攝圖檔之經緯度數據及時 間數據’以成為每一圖檔之檔名。其中時間數據單位可取至微秒。 一身型資料庫則係依據攝影機拍攝角度而建立的各種身型資料 庫。 一套格裝置則先將攝影機取得之每一禎動態物影像圖檔套上虛擬 格線’而該虛擬格線是以路徑上固定標記物長度或其他固定物長度為 參考而設置。 一身型比對裝置係將攝影機取得之一禎動態物全身影像圖檔，利 用影像比對方式’經與身型資料庫比對，判讀出該動態物之身型。 一速度計算裝置將所拍攝的至少兩禎以上圖檔相比對出距離差及 時間差，以計算出該動態物之速度。 7 200901099 其中身型資料庫、套格裝置、身型比對裝置及速度計算裝置係可 被設置於一後端伺服器中，而攝影機所拍攝之圖檔資料可經由有線戋 無線傳輸方式送至該後端伺服器，以進行辨識。 在第二圖中，係為本發明應用於公路上動態車輛之辨識系統之一 種實施例’其架構包括了第一攝影機、第二攝影機、一侧裝置、— 衛星定位系統（GPS)' -套線裝置、一車型資料庫、一車型比對裝置 及一車速計算裝置。200901099 IX. Description of the invention: [Technical field to which the invention pertains] The present invention relates to a dynamic object recognition system, and more particularly to an identification system for determining the body shape or quick material of a physical image file obtained by a camera. [Prior Art] On the A road, the vehicle model, the speed and the license plate _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ The image aging device has at least a camera and a flash, and when the radar is set to the vehicle overspeed, the image capturing device is controlled to be activated to obtain the overspeed vehicle photo data. When the conventional identification system is in use, since the speed measuring device transmits a light beam to the vehicle, such as laser or infrared rays, it is then possible to measure whether the vehicle is overspeeding. This allows the vehicle to be equipped with a detection machine to notify the driver in advance to avoid overspeed; this phenomenon makes the identification system more effective in preventing overspeed. Sometimes the 'image camera' must use the flash when shooting. This sudden flash has f " which can cause driving shock and cause traffic accidents. The traditional identification system needs to be improved. Due to the MJPEG (Moving Pictures Experts Group) technology, it is a dynamic image compression technology based on the static image compression technology JPEG. MJPEG can generate serialized moving images. The process does not consider the changes between different planes in the video stream. It only compresses each side separately. This compression method is very suitable for static pictures. The resolution can be from 352x240, 704x480. To 1280x1024. Therefore, the 5 200901099 continuous multi-complement images obtained from the camera will be clearly identifiable. The present invention will utilize this technique to improve the vehicle identification system. SUMMARY OF THE INVENTION The main object of the present invention is to provide a dynamic object recognition system, which uses a camera to set a photographic range from the top of the moving path, and when the dynamic object enters the photographic range, the device controls the device. The camera shoots a continuous image map building, and obtains the latitude and longitude data and time data of each icon via a satellite positioning system. (10)) to become the name of each floor. Each image file is first set with a virtual grid line by using a set-top device, and the virtual grid line is set with reference to the length of the fixed marker on the path or other fixed purity; the type comparison device will Take a full-body image file of the dynamic object obtained by f-iron, and use the image comparison method to touch the body shape of the secret. - The speed meter will compare the distance difference and the time difference between at least two or more images taken to calculate the speed of the dynamic object. In the present invention, since the shape and speed of the dynamic object are judged by a pure image, for example, in the judgment of the fire I, the vehicle on the road, the traveling vehicle on the road, and the sailing ship, it is only necessary to set up the camera at an appropriate position on the path of travel, and Change the environment and peripheral devices on the path. In particular, when the vehicle is used, the vehicle does not need a flash and emits a beam signal to identify the vehicle and the speed of the vehicle to reduce the chance of accidents caused by driving, and the identification system cannot be detected in the vehicle. According to the technology, the invention can also divide the body shape identification and speed identification into two independent independent systems to identify the body shape and speed respectively. Hereinafter, the system features and functions of the present invention will be described in detail based on the embodiments shown in the drawings. [Embodiment] Please refer to the first figure. The dynamic object identification system provided by the present invention can be used for trains, vehicles or boats, etc. The present invention mainly erects at least a suitable position on the path of the dynamic object. 'This camera sets the segment shooting range L. The shooting range L to v can capture the length of the whole body and the length of the segment to obtain a continuous image image of the Dolby High Resolution', for example, taken by MJPEG technology. A detecting device controls the camera to start shooting when it detects that the dynamic object enters the shooting range. A satellite positioning system (GPS) is used to obtain the latitude and longitude data and time data of each captured image to become the file name of each image file. The time data unit can be taken to microseconds. The one-size database is a collection of various body types based on camera shooting angles. The set of grid devices first sets each dynamic object image file obtained by the camera onto a virtual grid line ', and the virtual grid line is set with reference to the length of the fixed marker on the path or other fixture length. The one-piece comparison device obtains one of the dynamic body full-body image files of the camera, and uses the image comparison method to compare the body shape with the body type database to determine the body shape of the dynamic object. A speed calculation device compares the distance difference and the time difference between at least two or more captured images to calculate the velocity of the dynamic object. 7 200901099 The body type database, the checker device, the body type comparison device and the speed calculation device can be set in a back-end server, and the image data captured by the camera can be sent to the cable via wireless transmission. The backend server for identification. In the second figure, an embodiment of the invention is applied to an identification system for a dynamic vehicle on a highway. The architecture includes a first camera, a second camera, a side device, and a satellite positioning system (GPS). Line device, a model database, a model comparison device and a vehicle speed calculation device.
配合第三圖，本發明的第—及第二攝影機1A2以—架體^係被 木叹於公路4適當位置’例如高度H1為5 m，以Μ·影像格式拍攝 行進的車輛5影像；公路4上有車道分隔虛線6。 、第-攝影機1係'相對於公路4上設定—段至少超過最大車輔長度 以上及行進-段路㈣第—攝影細u，例如2q①，以取得經過該二 —攝影範圍L1内之車輛5影像之圖槽。 人 /而第—攝影機2設定_段第二攝影範圍U，例如6 m，該第 影範圍L2自第一& 影機拍攝叙歸！ 峨奴錢過車_被第一攝 以取侍紐過該第二攝影範圍内之車輛影像之圖檔。 偵測裝置係為控制用， —攝影機1拍攝車輛影像， 攝影機2拍攝車輛影像。 當車輛全身進入第一攝影範圍L1後起動第 且當車輛進入第二攝影範圍L2後起動第二 數據，以Π:: (GPS)用來取得每一拍攝圖檔之經緯度數據及時間 日13時24分16秒點3〇3微秒， 緯3八"、母—隨之槽名；例如GPS位置東經52度3分4秒，北 、、、毕38度7分6秒，時間2007年7月！ 8 200901099 則檔名為 52343876070701132416303. jpg。 料庫車型資料庫則係依據第一攝影機1拍攝角度而建立的各種車型資 套格裝置則先將第-及第二攝影機取得之每—禎車姉像圖播套 上虛擬格線’如第四圖所示’崎虛擬格線是以公路上固定標記物長 度為參考而設置，或者其他另行增加可供參相定物長度而設置例 如車道分隔纽6，或者另行以-已知長度之尺標為參考而設置。 車型比對裝置將第-攝影機i取得之車輛影像圖播，利用影像比 對方式’經與車黯料庫比對’綱出該車輛之車型。而車速計算裝 置，係將第二攝影範@ L2中取得之圖檔與第一攝影範@ u中取得之 圖檔比對，計算出該車輛之車速。 其中車型資料庫、套格裝置、車飢對裝置及車速計算裝置係可 被設置於-後端舰器中，而第—及第二攝影機所拍攝之圖檔資料可 經由有線或無線傳輸方式送至該後端舰^，崎行車辅辨識。 以較佳實施方式說明，此實施例中第一攝影機之攝影範圍為2〇 公尺，拍攝速度為30 frailies/秒，解析度為驗^，而第二攝影機 之攝影範圍為6公尺，拍攝速度為30 frames/秒，解析度為Half D1。 第-攝影機拍攝之圖檔套上虛擬橫格線間隔為2〇cm，共1〇〇條；第二 攝IV機拍攝之圖樓套上虛擬橫格線間隔為1〇⑽，共6〇條。經比對後， 即可精確計算出車速，及比對出車型。 由於第一及第一攝影機每秒拍設30禎（frames)，以MjpEG技術 特性，有車輛經過時為每禎間隔0.033秒，沒車輛經過時，第一及第 9 200901099 二攝影機不會回送影像圖片檔。 如以車輛之車柄_公里/㈣，車輛每〇. 〇33歸進距離為 2„ 78 m。而第二攝影機之第二攝影範圍L2為“，仍遠大於2】『 所以在此健實财式下，岐行進車輛之車勒公里/小時， 仍可被本發明辨識系統辨識出車速。 本發明由独純影像來崎公路上行進林之車型及車速，只需 在公路上適纽雜賴顯即可，不纽變姐上崎境及週邊設In conjunction with the third figure, the first and second cameras 1A2 of the present invention are photographed by a frame body at a suitable position of the road 4, for example, a height H1 of 5 m, and the image of the traveling vehicle 5 is captured in a video format; There is a lane separating the dotted line 6 on the 4th. The first camera 1 is set relative to the road 4 to at least exceed the maximum vehicle auxiliary length and the travel-segment road (4)-photographing fine u, for example, 2q1, to obtain the vehicle 5 passing through the second-photographing range L1. The image groove of the image. The person/the first camera 2 sets the second shooting range U, for example 6 m, which is taken from the first &峨 slave money over the car _ was taken by the first camera to take the image of the vehicle image in the second photographic range. The detecting device is for control, the camera 1 captures a vehicle image, and the camera 2 captures a vehicle image. When the vehicle body enters the first photographing range L1, the second data is started, and when the vehicle enters the second photographing range L2, the second data is started, and Π:: (GPS) is used to obtain the latitude and longitude data of each photographing image and the time of day is 13 o'clock. 24 minutes and 16 seconds, 3 〇 3 microseconds, latitude 3 eight ", mother - followed by the slot name; for example, GPS position east longitude 52 degrees 3 minutes and 4 seconds, north,, and 38 degrees 7 minutes and 6 seconds, time 2007 July! 8 200901099 The file name is 52343876070701132416303. jpg. The library model database is based on the shooting angle of the first camera 1 and the set of vehicle models is first set to the first and second cameras to capture the virtual grid line. The 'Saki virtual grid line shown in the four figures is set with reference to the length of the fixed marker on the road, or other additional lengths for the phase of the phase can be set, for example, the lane divider 6 or another known length. Set as a reference. The vehicle type comparison device broadcasts the vehicle image obtained by the first camera i, and uses the image comparison method 'by comparison with the vehicle magazine' to outline the vehicle type of the vehicle. The vehicle speed calculating device compares the image file obtained in the second photography model @L2 with the image file obtained in the first photography model @u, and calculates the vehicle speed of the vehicle. The vehicle model database, the box device, the vehicle hunger device and the vehicle speed calculation device can be disposed in the back-end ship, and the image data captured by the first and second cameras can be sent via wired or wireless transmission. To the back-end ship ^, Qishang car auxiliary identification. In a preferred embodiment, the first camera has a shooting range of 2 mm in the embodiment, a shooting speed of 30 frailies/second, and a resolution of 6 meters. The second camera has a shooting range of 6 meters. The speed is 30 frames/second and the resolution is Half D1. The interval between the virtual camera lines on the first camera is 2〇cm, which is 1〇〇; the interval between the virtual camera lines on the second camera is 1〇(10), a total of 6 . After comparison, the speed can be accurately calculated and the vehicle can be compared. Since the first and first cameras are set to 30 frames per second, the MjpEG technical characteristics are 0.033 seconds per vehicle when passing the vehicle. The first and the 9th 200901099 cameras will not return images when there is no vehicle passing. Image file. For example, the vehicle handle _ km / (four), the vehicle per 〇. 〇33 return distance is 2 „ 78 m. And the second camera's second photographic range L2 is “, still far greater than 2 】” So here is the health Under the financial model, the speed of the vehicle and the hour of the vehicle can still be recognized by the identification system of the invention. The invention adopts the purely pure image to drive the model and the speed of the forest on the road to the road, and only needs to be displayed on the road, and it is not necessary to change the appearance of the sister and the surrounding area.
備，夜_另輔脉外_明設備投注在目魏補足全天候⑵隱的 紅外線光源即可。尤其’不需賴光燈，喊少駕較驚而發生意外 之機會。 進者’本發明在辨識行進車輛之車速時，不需發出雷射或紅外線 專光束㈣，使縣_無法細職觸线。 改變，將可達到辨識到車輛極高的速度，極為實用。 據又疋 在以上·純影像來判斷動態物之身型及速度之方法，也可分別 應用於-動態物身型辨識系統（如第五圖所示）及_動態物速度辨識 系統（如第七圖所示）。 參見第五® ’係為-種祕物身型辨齡統，具有_攝影機拍攝 MJPEG影雜紅_，#設置崎_徑上適纽置，姆於該行經 路徑上設1段至少超着祕全身長度賴f彡麵，㈣得經過該 攝影範Μ之動祕影像之關。而其他侧裝置、魅定位系統、 套格裝置、身型·庫及身觀雌置#，職觀綱。如此，以 構成一獨立的動態物身型辨識系統。 200901099 第八圖係為該系統應用在公路上動態車輛車型辨識的實際配置 圖’係在公路4上適當位置設置攝影機11，該攝影機11設定了- 範圍L3以拍攝經過的車輛；進而依據本發明圖播之提供檔名、 套格、比對等，而可觸細之車型。 '見第七圖’係為—種麟物速度辨識系統，具有-攝影機拍攝 Μ罵影絲式之_，概侧满虹適纽置奴—段攝影 範圍’該攝_圍至少可取得兩福以上經過該攝影範_之動態物影 * €❿其他動彳裝置、衛星定赠、統、套格裝置及速度計算裝 置等職前述相同。如此，以構成一獨立的動態物速度辨識系統。 第八圖係為該系統應用在公路上動態車輛車速辨識的實際配置 圖’係在公路4上適當位置設置攝影機21，該攝影機21設定了一 •k攝’5V範圍L4以拍攝經過的車輛；轴依據本發明賴之提供權名、 套格、速度#料，巾可辨識車輛之車速。 顯然地本發明並不以實施方式中所例實施例為限，其他依據本 發月技HX嶋^像來判斷行進車輛之各種可能之資料，如為等效 之«變化’例如攝影機以其他高解晰度影像格式拍攝，或攝 -架或兩架社等，仍屬本案補倾細之中。 ” 【圖式簡單說明】 第一圖代表本發明之系統架構圖， 第-圖代表本發明應於公路上動態車輛辨識系統之架構圖， 第三圖代表本發明攝影機震置於公路上及設定攝影範圍之實作示 11 200901099 意圖， 第四圖代表本發贿縣置糊财上倾格叙示意圖， 第五圖代表本發明之動祕身型架翻， 於公路上及設定攝影範圍之實作示意 第六圖代絲發贿祕姆林讀影 圖 第七圖代表本發明之動態物速度辨識系統架構圖， r 第八圖代表本發明應用於公路上車輛車速 於公路上及設定攝影範圍之實作示意圖。 5 〃、之影機裝置 【主要元件符號說明】 第一及第二攝影機……1及2 架體......3 公路......4 車輛......5 車道分隔虛線......6。 攝影機......11、21 12Preparation, night _ other auxiliary pulse outside _ Ming equipment betting in the eye to complement the full-time (2) hidden infrared light source. In particular, there is no need to rely on the light, and the chance of an accident is unexpected. In the present invention, when identifying the speed of a traveling vehicle, it is not necessary to emit a laser or infrared special beam (4), so that the county cannot contact the line. The change will be able to reach the extremely high speed of the vehicle and is extremely practical. According to the above method, the method of judging the body shape and speed of the dynamic object can also be applied to the dynamic body type identification system (as shown in the fifth figure) and the dynamic object speed identification system (such as Figure 7). See the fifth ® 'system for the type of secret body type, with _ camera shooting MJPEG shadow red _, # setting 崎 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ The length of the whole body depends on the face, and (4) has to go through the secret image of the photography. The other side devices, the charm positioning system, the set of devices, the body type · library and the body of the female set #, job view. In this way, an independent dynamic body type identification system is constructed. 200901099 The eighth picture shows the actual configuration map of the system for dynamic vehicle model identification on the road. The camera 11 is set at an appropriate position on the road 4, and the camera 11 sets the range L3 to photograph the passing vehicle; The map broadcast provides the file name, the box, the comparison, etc., and the model can be touched. 'See the seventh picture' is a kind of forest speed identification system, with - camera shooting Μ骂 丝 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ The above is the same as the above-mentioned dynamic image*, other mobile devices, satellites, satellites, kits, and speed calculation devices. Thus, an independent dynamic object velocity recognition system is constructed. The eighth figure is the actual configuration diagram of the system for applying the speed identification of the dynamic vehicle on the road. The camera 21 is set at an appropriate position on the road 4, and the camera 21 sets a range of L4 of the '5-photograph '5V to photograph the passing vehicle; According to the invention, the axis can provide the right name, the set, the speed, and the towel can identify the speed of the vehicle. It is obvious that the present invention is not limited to the embodiments exemplified in the embodiments, and other various materials according to the present invention are used to judge various possible materials of the traveling vehicle, such as equivalent "changes" such as other high resolution of the camera. The shooting of the image format, or the photo frame or the two agencies, is still in the case of this case. BRIEF DESCRIPTION OF THE DRAWINGS The first figure represents the system architecture diagram of the present invention, the first diagram represents the architecture diagram of the dynamic vehicle identification system on the road, and the third diagram represents the camera of the invention being placed on the highway and set. The actual scope of photography 11 200901099 Intent, the fourth picture represents the schematic diagram of the bribe county, and the fifth picture represents the moving body of the invention, on the road and setting the scope of photography. The seventh figure represents the architecture of the dynamic object speed identification system of the present invention, and the eighth figure represents the application of the present invention to the speed of the vehicle on the road and the setting of the photographic range. Schematic diagram of the implementation. 5 〃, the camera device [main component symbol description] First and second camera ... 1 and 2 frame ... 3 road ... 4 vehicles ... ...5 lane separation dotted line...6. Camera...11,21 12