200904673 九、發明說明： 【發明所屬之技術領域】 本發明涉及一種車輛防撞系統及其方法，尤其是一種光 學影像車輛防撞糸統及其方法。 【先前技術】 ; 習用之車輛防撞系統係利用至少一個雷達或固定式攝影 機’以測知車輛前方是否有障礙物，以及車輛與前方車輛的距 離’當系統發現於車輛前方一安全距離内出現障礙物時，則發 出警告以提醒駕駛迴避障礙。 然而’為了制有效測量之目的，上述之雷達常會產生 過高功率之電磁波，因而造成人體的不適，抑或造成生理上長 i A之後遺症。若降低電磁波功帛，即會造成測量距離銳減，使 得車輛防撞系統之功用大幅降低，並且使用雷達預警無法提供 比影像更直接且完整之道路資訊。 另外，上述之固定式攝影機雖然可提供較完整之道路資 訊’但卻需經由獅之幾何演算崎&實際之安全轉，而此 -複雜運算即令H車輛防齡統難以即時提供車輛前方 之訊息’大大降低此-車輛隨纽之膽舰。再者，固定 式攝影機必會有其影像捕捉之死角，因而減少傳統之車輛防撞 5 200904673 系統的預警準確度 更因此危害了車輛行使安全。 f發明内容j 之需 繁於上述之翻f景巾，為 求’本發明提供 職上k利& 述傳 種先予衫像車輛防撞系統可用以解決上 統之車輛防衫絲能達成之標的。 法。利肢 係提供—種絲影料_系統與方 送至影賴轉置偵測車輛顯影像，並將該影像 =一4_單元進行輯I作觸程序。若 :定=像為-障礙物，該影像辦別單元傳送-警告二 。讀置，以提供駕駛前方障礙物訊息。 根據X上所述之目的，此—光學影像車輛防㈣統 :速度感測裝置、至少-影像擷取裝置及-影像辦別單元。藉 取传則方障礙物資訊。 【實施方式】 本發明在此所探討的方向為一種光學影像車辆防撞系 統。為了能徹底地瞭解本發明’將在下列的描述中提出詳盡的 方法及其結構。顯舰，本發_施行並被定於光學景^車 200904673 輛防撞系統之技藝者所熟習的特殊細節。另一方面，眾所周知 的結構或方法並未描述於細節中，以避免造成本發明不必要之 限制。本發明的較佳實施例會詳細描述如下，然而除了這也詳 細描述之外，本發明還可以廣泛地施行在其他的實施例中，且 本發明的範圍不受限定，其以之後的專利範圍為準。 參考第-圖所示’本發明係提供一種光學影像車輛防撞 系統100以幫助駕駛人更能輕鬆地瞭解車輛四周駕駛環境的 安全與否’其中上述之光學影像車輛防撞系統100係藉由一速 度感測裝置102以感測一機動載具108之速度，並藉由至少一 影像擷取裝置104擷取連續之複數張晝面11〇，其中連續之複 數張晝面110係可為機動載具108前方之連續影像，且至少一 影像擷取裝置104係根據機動载具1〇8之速度調整俯角， 如第二圖所示，上述之至卜影像擷取裝置1G4之俯角ιΐ2隨 著機動載具108之速度增加而減少，亦隨著機動载具之速 度減少而增加。 另，上述之光學影像車輛防撞系統1〇〇更藉由_影像辦 別單元106對複數張畫面110進行一影像辨識程序，並且在影 像辨識程序中辨識出一警示圖形1H時輸出-警告訊號116" 藉此以警告駕駛人。其中更可藉由—警示I置118接收此一馨 告訊號116並發出-警示訊息12〇，例如藉由一飢接收警^ 訊號II6並發出-料聲響，或是勤—閃光裝_收警告訊 200904673 號116並發出一警示閃光。 其中上述之影像辨識 ‘所預測出之動像辨識出上述 之警示圖形 序包含一動像預測程序，此一影 像辨識程序魏據動像_程序： 一 序係了於複數張畫面110中界定一馨 不區，當動像預測裎序所預 。 接觸時，Mm $之動像自警示區外並向警示區 接觸時4胸_可辨物示圖形114。 例如動像到程序在複數張畫面⑽之中央標定一橫向 2即向上細下各延伸—特定距離，動像預測 疋叫向界線騎心之特定範圍係為上述之邀 不區:當動像預測程序所預測出之動像係位於警示區之上方並 與警示區接觸時（亦即當物辦 _ w h㈣由警碰之上方接近並接觸警示 影像辨識程序即辨識出警示圖形m。除了上述實施 例之外’亦可藉由上述之影像辨識程序辨識出地面之兩條車道 線之影像轉定出警祕之_，並且麵區亦包含—馨戒線 以界定出警示區之上側，#動像預_序所制出之動像自警 不區外亚向警示區接觸時（亦即當物體自警示區之上方向警示 區接觸，或是物體自警輯之兩側向兩條車道線接觸時）= 像辨識程序即可辨識出警示圖形m。除了上述方法以外，本 發明更揭露了頂實施例以達成車輛防撞之目的。 200904673 例如，當上述之動像在一晝面中 限值時，影像_程序即辨識出警之=例超出一門檀 _。其令上述之動像係可為-障礙圖=隨即輸出警告訊 輛等等）出現於畫面中，當此一障礙物^硕、動物或車 門檻限值時，即會被辨識為上 j例大於上述之 物過大時，警告訊請便恤_114,亦即當障礙 所產述讀像村藉树續複數張畫關的差異處 所產生，例如某一張書面並未 可僅為草… 出車輛，而隨後幾張晝面(亦 ，為某一張晝面之下一張）卻出現車輛，亦即判斷車輛係為 上述之警不圖形114，而書 …、、、 如藉由财輛之—料即可，例 斷盆為:I車輛之車牌、後車燈、排氣管或後輪即可判 _外，僅藉由單―—^複數張畫面以辨識警示圖 此皆為孰幻Γ 輕易辨識出警示圖形114， 贅述：…關領域者所能輕易推知之先前技術，在此不加以 再者，上叙影像顺程耗翻事細識紐面之車 m射賴㈣鱗道狀影細狀-車道，且 此-車道係位於兩條車道線之間。 與位m：r示圖形m區分為位於車道内 更辨識。例如當車輛（警示圖形114)出 200904673 現於車勒時’ _發料告訊號116告知駕駛人，藉此以降 低車速或閃避前方車輛’亦可控制機動載具應自動降速直到 上述之前方車輛_畫面警示區。若車輛出現於車道外時，及 藉由判斷車輛是顿兩條車道線之影像之—树，當車輛與兩 條車道線之影像之-姆時，即發丨警告峨⑽告知 以閃避突然插入前方車道之車輛。 其中上述之影像掏取裝置1〇4為下列族群之一或其組 合·’感光耗合元件影像摘取裝置、互補式金屬_氧化層_半導體 影像擷取裝置及類比影像摘取裝置。 上述俯角112係依據下式求得： a = tan1 X/L， 其中α為上述俯角112， L為-預設距離’此—職轉係根據機賴具之速度調整， 當速度越快，預設距離必然越長，反之亦然。 X為至少一影像擷取裝置104距地面高度。 本發明第二實施例係揭露一種光學影像車輛防撞方法， 上述光學影像車輛防撞方法包含下列步驟： 感測一機動載具之速度； 設置於機動載具之至少一影像擷取裝置根據機動載 具之速度調整至少一影像擷取裝置之俯角； 至少一影像擷取裝置擷取連續之複數張畫面；以及 200904673 叹狀機動财之—影像辦料域 t影像_料，纽在影_識辦巾_出:= 圖形時輸出一警告訊號。 ⑺ 像辨=像觸料包含—崎制辦，其中影 圖形。當動像^像預測程序所預測出之動像辨識出警示 超出〜田_值貝時程序所預測出之動像在一影像中所佔之比例 之動像=Γ影像辨識程序辨識出警示圖形。其中上述 異處，亦可It由複數張影像之其中—張所辨識出。 像，識—影 道線之旦增Γί 界定出—車道，車道位於兩條車 吉p 0 w警相職位於車勒時 =直至警缩糊概咖。_=3 其他部分。5不圖形係一車輛之車牌影像或可為-料之 張書面除法「之外’上述之動像預測程序更可於複數 面令界疋-警不區’當動像預測程序所預測出之動像自盤 —時，影像辨識程序即 圖 形，其中上述之影像辨顧序辨識舰面之兩條車 200904673 以界疋出警示區之兩側，並且警祕包含—警戒線以界定出警 不區之上侧。另外，上述之警示區係可位於複數張晝面之中央。 另，上述之至少一影像擷取裝置之俯角隨著機動載具之 絲增加而減少’亦隨著機動載具之速度減少而增加。且更可 藉由—警示裝置以接收警告訊號並發出一警示訊息。BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle collision avoidance system and a method thereof, and more particularly to an optical image vehicle collision avoidance system and method therefor. [Prior Art]; A conventional vehicle collision avoidance system utilizes at least one radar or fixed camera 'to detect whether there is an obstacle in front of the vehicle and the distance between the vehicle and the preceding vehicle' when the system is found within a safe distance in front of the vehicle When there is an obstacle, a warning is issued to remind the driver to avoid the obstacle. However, for the purpose of making effective measurements, the above-mentioned radars often generate electromagnetic waves that are too high in power, thereby causing discomfort to the human body or causing physiological sequelae. If the electromagnetic wave function is reduced, the measurement distance will be sharply reduced, so that the function of the vehicle collision avoidance system is greatly reduced, and the use of radar warning cannot provide more direct and complete road information than the image. In addition, although the above-mentioned fixed camera can provide more complete road information, it needs to be calculated by the lion's geometry to calculate the actual safety of the squid, and this complex operation makes it difficult for the H vehicle anti-age system to provide the information in front of the vehicle. 'Largely reduce this - the vehicle with the New Zealand. Moreover, the fixed camera will have its own image capturing dead angle, thus reducing the traditional vehicle collision avoidance. 5 200904673 The warning accuracy of the system is more harmful to the safety of the vehicle. f invention content j need to be versatile in the above-mentioned f-fake towel, in order to seek the 'invention of the present invention, the k-language and the pre-existing type of pre-emptive vehicle anti-collision system can be used to solve the upper body of the vehicle anti-shirt wire can be achieved The subject matter. law. The limbs provide a kind of silk shadow material _ system and party to the film to detect the vehicle display image, and the image = a 4_ unit for the I touch program. If : = = like - obstacle, the image device unit transmits - warning two. Read to provide an obstacle message in front of the driver. According to the purpose described in X, the optical image vehicle defense system is: a speed sensing device, at least an image capturing device and an image processing unit. Borrow the information on the obstacles. [Embodiment] The invention discussed herein is an optical image vehicle collision avoidance system. In order to fully understand the present invention, detailed methods and structures thereof will be set forth in the following description. The display ship, the hair _ implementation and is set in the optical Jing ^ car 200904673 anti-collision system is familiar with the special details of the craftsman. On the other hand, well-known structures or methods are not described in detail to avoid unnecessarily limiting the invention. The preferred embodiments of the present invention will be described in detail below, but the present invention may be widely practiced in other embodiments, and the scope of the present invention is not limited, and the scope of the following patents is quasi. Referring to the figure - the present invention provides an optical imaging vehicle collision avoidance system 100 to help a driver more easily understand the safety of the driving environment around the vehicle. The optical image vehicle collision avoidance system 100 described above is used by the present invention. A speed sensing device 102 senses the speed of a motorized vehicle 108 and draws a plurality of consecutive planes 11〇 by at least one image capturing device 104, wherein the plurality of consecutive sheets 110 are maneuverable a continuous image in front of the carrier 108, and at least one image capturing device 104 adjusts the depression angle according to the speed of the motorized vehicle 1〇8. As shown in the second figure, the above-mentioned image capturing device 1G4 has a depression angle ιΐ2 The speed of the mobile vehicle 108 increases and decreases, and also increases as the speed of the mobile vehicle decreases. In addition, the optical image vehicle collision avoidance system 1 described above further performs an image recognition process on the plurality of pictures 110 by the image processing unit 106, and outputs an alarm-warning signal when a warning pattern 1H is recognized in the image recognition program. 116" Use this to warn the driver. In addition, the warning indicator 116 can receive the stimuli signal 116 and issue a warning message 12 〇, for example, by receiving a warning signal from the hunger, and outputting a sound, or a flashing light. Newsletter number 210904673 and issue a warning flash. The above-mentioned image recognition 'the predicted moving image recognizes that the warning graphic sequence includes a moving image prediction program, and the image recognition program is based on the moving image_program: a sequence is defined in the plurality of pictures 110 to define a sweet No district, when the motion prediction is pre-ordered. When in contact, the movement of Mm $ is from the outside of the warning zone and to the warning zone. For example, the moving image to the program is calibrated in the center of the plurality of pictures (10) with a horizontal direction 2, that is, an upward extension and a specific distance, and the specific range of the motion picture prediction squeaking to the boundary line is the above-mentioned invitation area: when the motion picture is predicted When the motion picture predicted by the program is located above the warning area and is in contact with the warning area (ie, when the object office _ w h (4) is approached by the upper side of the warning and contacts the warning image recognition program, the warning pattern m is recognized. In addition to the above example, the image of the two lane lines on the ground can be identified by the image recognition program described above, and the area is also included - the ring line is defined to define the upper side of the warning area, #moving image The motion picture produced by the pre-sequence is not in contact with the warning area outside the police zone (that is, when the object touches the warning zone from above the warning zone, or when the object comes into contact with the two lane lines from both sides of the police series) = The warning pattern m can be recognized by the recognition program. In addition to the above method, the present invention further discloses the top embodiment for the purpose of achieving vehicle collision avoidance. 200904673 For example, when the above-mentioned moving image is at a mid-limit value, Shadow _ The program recognizes the police = the case is beyond a door _. It causes the above-mentioned motion picture system to be - the obstacle map = the output warning message, etc.) appears in the picture, when the obstacle is a master, animal or car When the threshold value is reached, it will be recognized as the upper j case is larger than the above-mentioned thing, the warning message is _114, that is, when the obstacle is produced by the village, the difference between the image and the tree is repeated. For example, if a certain piece of paper is written, it is not only for the vehicle to exit the vehicle, but then several vehicles (also, one below the surface of one of the faces) are present, that is, the vehicle is judged to be the above-mentioned police. 114, and the book...,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, -^Multiple pictures to identify the warning maps are all illusions 轻易 Easy to identify the warning graphic 114, narration: ... the prior art that can be easily inferred by the domain, no longer Things are familiar with the car of the noodle (m), the scales of the scales, the lanes, and the lanes are located on two lanes. Room. It is distinguished from the position m: r to indicate that the graphic m is located in the lane. For example, when the vehicle (warning pattern 114) is out of 200904673, the vehicle sends a warning message 116 to the driver to reduce the speed of the vehicle or to avoid the vehicle in front, and the motor vehicle should be controlled to automatically slow down until the preceding party. Vehicle_screen warning area. If the vehicle appears outside the lane, and by judging that the vehicle is the image of the two lane lines, when the vehicle and the image of the two lane lines are the same, then the warning 峨(10) informs that the ducking is suddenly inserted. The vehicle in the front lane. The image capturing device 1〇4 is one of the following groups or a combination thereof, a photosensitive sensing device image capturing device, a complementary metal oxide layer semiconductor image capturing device, and an analog image capturing device. The above-mentioned depression angle 112 is obtained according to the following formula: a = tan1 X/L, where α is the above-mentioned depression angle 112, L is - the preset distance 'this--the rotation is adjusted according to the speed of the machine, and the faster the speed, the pre- Set the distance must be longer, and vice versa. X is the height of at least one image capturing device 104 from the ground. A second embodiment of the present invention discloses an optical image vehicle collision avoidance method. The optical image vehicle collision avoidance method includes the following steps: sensing a speed of a mobile vehicle; and at least one image capturing device disposed on the mobile vehicle according to the maneuver The speed of the vehicle is adjusted to at least one depression angle of the image capturing device; at least one image capturing device captures a plurality of consecutive images; and the 200904673 sighs the mobile money-image storage area t image_material, New in the shadow_ Take a towel _ out: = a warning signal is output when the graphic is displayed. (7) Image recognition = Image contains - Kawasaki Office, which is a graphic. When the moving image predicted by the moving image prediction program recognizes that the warning exceeds the value of the image predicted by the program in the image in the image = the image recognition program recognizes the warning graphic . Among the above, the difference may be recognized by one of the plurality of images. Like, the sensation-shadow line is increased Γ 界定 Defining the lane, the lane is located in two cars, p 0 w, the police position in the car, = until the police. _=3 Other parts. 5 No graphic is a license plate image of a vehicle or can be a written division of the material. "Besides the above-mentioned moving image prediction program can be used in the multiple-faced---------------------------- When the video is self-discussed, the image recognition program is a graphic, in which the above-mentioned images identify the two vehicles 200904673 of the ship's surface to define the two sides of the warning zone, and the police contain a warning line to define the police zone. In addition, the above-mentioned warning zone may be located at the center of the plurality of facets. In addition, the depression angle of at least one of the image capturing devices described above decreases as the wire of the mobile vehicle increases. The speed is increased and increased, and the warning device can be used to receive the warning signal and issue a warning message.
Example 1 智慧型道路運輸系統（ITS，Intdligent Transp〇rtati〇n :m)係藉由先進之電腦、資訊、電子、通訊與感測等科技 〜用’透過所提供即時資訊的溝通與連結，以改盖人、車、 間的互動關係，進而增進運輸系統一 制通環境衝擊。本發明藉由輸入-速度限 該安全距全雜，钱設有本判之車減與前車保持 迷鱼用此一特性’本發明可進-步測出前方車輛之 -:成本遞這些訊息，本發明可建構出 構造早純之智慧型道路運輪系統。Example 1 The Intelligent Road Transport System (ITS, Intdligent Transp〇rtati〇n:m) uses advanced computer, information, electronics, communication and sensing technologies to communicate and link through the instant information provided. Change the interaction between people, cars and houses, and then improve the environmental impact of the transportation system. According to the invention, the safety distance is completely mixed by the input-speed limit, and the money is provided with the vehicle reduction of the present judgment and the maintenance of the front vehicle to use the fish. The invention can further measure the vehicle in front - the cost of delivering the message The invention can construct an intelligent road transport system which is constructed as early as pure.
Example 2 測到事故^像^=統可即_取車Μ靖4，當系統偵 _史備，Γ記錄將立即啟動，運用既有的影像 中。透的影像做壓縮處理咖 事_練’有利於事後對料事現場的重建與責 12 200904673 任之髮清工作。同時，更可以連結無線通訊系統，在第一時間， 主動聯繫相關警消單位，以加速人員的援救與事故的處理速 度。本發啊裝設至少—影像擷取裝置於車婦後左右，涵蓋 車輛顺360度之並連續記雜觸取之轉，形成一 行車影像記錄系統以幫助事故發生時肇事責任之餐清。 Example 3 如第—圖所不’駕駛人在裝設有傳統道路交通號誌、300 的路段’若想從傳統道路麟取得導路龍，需分散其對於路 況的注意力觸賴交通齡。本發明具有至少—影像操取裝 置及-影像_單元，若預先蚊字歧二維條碼方式於特定 路面上代表狀資訊之色塊，如車速關、道路名稱。當 車輛订驶至該歡路面，齡至少—影細取裝置取得路面標 記’並利㈣像_與運算單元_該標記所絲之資訊，以 幫助駕驶可於專注路面狀況的娜下取得翁纽，增進行車 安全。Example 2 Detecting an accident ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Translating the image to do the compression processing of the coffee _ _ is conducive to the reconstruction and responsibility of the scene after the event 12 200904673 Ren Zhiqing work. At the same time, it is possible to connect to the wireless communication system and, at the first time, actively contact the relevant police units to speed up the rescue and accident handling of personnel. This hair is equipped with at least the image capturing device behind the woman, covering the vehicle in a 360-degree direction and continuously recording the touch and turning, forming a driving image recording system to help the accident when the accident occurs. Example 3 If the driver does not have a traditional road traffic sign, 300 roads, if you want to get a guide from the traditional road, you need to distract the attention of the road to the traffic age. The present invention has at least an image manipulation device and an image_unit. If the two-dimensional barcode method is used in advance, the color block representing the information on a specific road surface, such as a vehicle speed and a road name. When the vehicle is scheduled to drive to the road, the age is at least - the shadow taking device obtains the road marking 'and the information of the _ and the arithmetic unit _ the mark to help drive the driver can concentrate on the road conditions Increase the safety of the car.
Example 4 先進安全車輛(ASV，Advanced Safety Vehicle)是應用資 訊電子材料專各種領域的先進技術’配合車輛的操縱控制， 以提昇車輛行車安全性’而其中完全不需要駕駛者操控，而能 夠確保行車安全的自動車輛駕驶系統為其核心主幹。本發明運 200904673 用其車輛防撞功能，並配合其它感測骏置，如入球 (GPS, Global Positioning System) . ^ ^ ^ 備，實現一自動車輛駕駛系統。 叹 顯然地’賴上面實闕，本發啊 的修正與差異。因此需要在其附加的權利要求項 2 ^例^雜的崎，本侧⑽泛地在其他的 〜1把仃上述僅為本發明之較佳實施例而已，並非用以 ^疋本發明之”專利範圍；凡其它未脫離本發明所揭示之於 Γ所完成轉效轉_飾，触包含在下獅請專利範^ 14 200904673 【圖式簡單說明】 第一圖係為光學影像車輛防撞系統結構示意圖； 第二圖係為車速變化與俯角大小關係示意圖；以及 第三圖係為道路資訊自動取得系統示意圖。 【主要元件符號說明】 100 光學影像車輛防撞系統 102 速度感測裝置 104 至少一影像擷取裝置 106 影像辦別單元 108 機動載具 110 複數張晝面 112 俯角 114 警示圖形 116 警告訊號 118 警示裝置 120 警示訊息 300 傳統道路交通號誌 15Example 4 Advanced Safety Vehicle (ASV) is an advanced technology in various fields of information electronic materials. It is used in conjunction with vehicle control to improve vehicle safety. It does not require driver control at all, and can ensure driving. A safe automatic vehicle driving system is its core backbone. The invention of the invention 200904673 uses its vehicle anti-collision function, and cooperates with other sensing, such as the GPS (Global Positioning System) to implement an automatic vehicle driving system. Sighing obviously, the remedy and difference between the above and the original. Therefore, in the appended claims, the present invention is not limited to the preferred embodiment of the present invention. The scope of patents; other than the disclosure of the invention disclosed in the 转 完成 完成 , , , , , , , , , , , , , , 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 The second diagram is a schematic diagram of the relationship between the vehicle speed change and the depression angle; and the third diagram is a schematic diagram of the road information automatic acquisition system. [Main component symbol description] 100 optical image vehicle collision avoidance system 102 speed sensing device 104 at least one image Capture device 106 Image processing unit 108 Mobile vehicle 110 Multiple sheets 112 Depression angle 114 Warning graphic 116 Warning signal 118 Warning device 120 Warning message 300 Traditional road traffic number 15