TW200922816A - Method and device for detecting the lane deviation of vehicle - Google Patents
Method and device for detecting the lane deviation of vehicle Download PDFInfo
- Publication number
- TW200922816A TW200922816A TW096145498A TW96145498A TW200922816A TW 200922816 A TW200922816 A TW 200922816A TW 096145498 A TW096145498 A TW 096145498A TW 96145498 A TW96145498 A TW 96145498A TW 200922816 A TW200922816 A TW 200922816A
- Authority
- TW
- Taiwan
- Prior art keywords
- lane
- vehicle
- image
- lane line
- line
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000000694 effects Effects 0.000 claims abstract description 12
- 238000012545 processing Methods 0.000 claims abstract description 9
- 230000000007 visual effect Effects 0.000 claims abstract 2
- 238000001514 detection method Methods 0.000 claims description 20
- 230000008569 process Effects 0.000 claims description 18
- 238000006073 displacement reaction Methods 0.000 claims description 12
- 238000012937 correction Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000012360 testing method Methods 0.000 claims description 3
- 241000886545 Leptodea leptodon Species 0.000 claims description 2
- 230000008859 change Effects 0.000 claims description 2
- 238000005192 partition Methods 0.000 claims description 2
- 238000012546 transfer Methods 0.000 claims description 2
- 206010011224 Cough Diseases 0.000 claims 1
- 208000001613 Gambling Diseases 0.000 claims 1
- 238000000926 separation method Methods 0.000 claims 1
- 239000012634 fragment Substances 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 206010039203 Road traffic accident Diseases 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 1
- 241001494479 Pecora Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035484 reaction time Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Landscapes
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
Description
200922816 九、發明說明: 【發明所屬之技術領域】 本發明係有關-種車輛偏移之檢知方法與裝置,詳而言之係、一種利用 影像判斷車祕,並_二次轉擬合方程絲取車麵移舞量及車道 曲率,藉此判斷行車是否正常並可提供警示之功效。 【先前技術】 隨著車_普及廣泛應用於人類社會,使得各地差距不斷縮短,逐漸成 為全球化-體之社會形態,但所謂—正必有—反,近幾年因交通事故傷亡 事件層出不窮先賴家及各A車辆飾尋求如何紐技應用於車 輔上,以私:尚車輛行駛的安全性,並降低交通事故發生率; 交通事故的發生,往往人為因素佔大部分,疲勞與分神則往往是意外 發生的主因,雖然國内外相關機構皆致力於瞌睡偵測、腦電波等相關研究 以提醒駕駛者,但效果卻差強人意,加上线反應時間及整體成本,無法 全面普及; 以往國内科技,過去均使用超音波雷達或光反射等物理手段達到車體 防護警訊效果;如中華民國專利申請案號第09221746 0號「可避 免駕駛人偏離車道之警告裝置」,其主要係於車輛之兩側各裝設一車輛警告 震置;該車輛警告裝置其内設有一電壓電源供應器、一光量感應器、一訊 號放大器、一電壓比較放大器及一警告顯示器所組合而成;其中,藉由該 光量感應器感應道路分隔線反射之光量,並利用道路分隔線所反射之光量 大於道路沒有分隔線部份,使該光量感應器藉由反射之光量差異而產生不 200922816 同的電流值,該電流值經過訊號放大器的調變,將電流值轉調變成適合之 電壓差’再以電壓比較放大器進行電壓比較放大,並得到一電壓輪出訊號, 再將電壓輸出訊號送至警告顯示^,並藉由該警告齡器提醒駕歇人,用 以提醒駕駛人所駕駛之車輛已偏離行駛中的車道; 近年來’由於軟體撰寫之彈性、c cD、CM0 s攝影機架設之便利、 所擷取資料多元化、成本較低紐勢,提供—全新的技術發展方向,因此 搭配C C D、CMO S攝織之高效率影像處理技術正勒發展,除了倒 車影像顯科,影像處理祕更可躲其他廣泛之安全防賴助,如偵測 車輛别方道路訊息,估算車輛目前行駛車道植置等等;如巾華民國專利 申請案號第0 9 213 4 0 9 5號「車道偏移警示裝置與方法」其包括一 車道視訊資訊擷取單元、-車道偏移侧單元、和__警示單元;其判斷車 道偏移的方法包含影像輸人、影像分析處理、搜尋輯設定、車道偏移憤 測、以及能自級動或關警示車道偏移的步驟,自動細車輛不當偏移 車道的情形’並給予駕駛人警示轉障行車安全,更包括—個自動判斷機 制決定暫時_警示裝置,以免錯誤動作頻繁而增加駕歇人困擾; 然而此種制技術係以道路標線與行車路徑進行崎,以確認車輛是 否產生偏移正f軌道;但此種彻比對方式並非萬無—失,仍有比對上差 異造成錯誤判斷之可能’此外,實際道路標線與行車路徑兩者並未經過轨 跡預估及運算,料產生實f偏差,發送錯騎訊,妓太晚發送警訊, 不利於實質使用; 是以,針對上述習知技術所存在之問題點,如何開發—種更具理想實 200922816 用性之靖結構,實使用消費者所殷切企盼,亦係相關業者須努力研發突 破之目標及方向。 有鑑於此,本發明人基於多年從事相關產品之製造開發與設計經驗, 針對上述之目標,詳加設計與審慎評估後,終得一確具實用性之本創作 【發明内容】 本發明之主要目的録純供—種賴偏移之檢知方法缝置,其可 改善習用技較限於制_較短,造成反應_不足,甚至由於比對方 式不佳,產生誤判、發送錯誤訊號,不利於實際使用等缺失。 為達到前述之發明目的,本發明之裝置係包括攝像單元'運算平台、 信號輸出單元,而其方法係包括影像辨識處理程序以及偏離估算處理程 序,其中: 土 該影像辨識處理轉係包括下列步驟。影細取步驟:攝像單元 祿取車輛前方道路影像晝面資料;車道線辨識步驟:將影像畫面資料= 為上半部以及下半部,像晝面資料之上半部侧以進行日夜間判斷刀 該影像畫面資料之下半部係再分為上下二部份,影像畫面資料距離車輛較 近端之部份轉為較低解析度,距離車輛較遠端之部份維持原解析度丨—辨 識流程,包括三種車道線辨識方式:高灰階值辨識、車道線邊緣特性辨識、 車道線寬度辨識,必須同時符合上述三細識方式,才得以判定為正確之 車道線; 該影像畫面資料下半部係由下社,劃分為若干等分區間並由下而 上進行下列流程:起始點搜尋流程:搜尋兩側車道線之起始點作為起點; 200922816 侧車道線流程:在目前區間内,利用上述辨識流程判斷出車道線,並利 用二次擬合曲線方程式進行車道線趨勢預估,連接所有片段之車道線做出 二次曲線;财修正雜:_ R 0 1侧修正車道_勢難實際靜. 如此重錢行上錢程,以____,錢車魏預測模型 不斷修正擬合實際車道線軌跡;又,藉由連續數張影像晝面資料判定消失 點位置是鶴近或_,明加狀準確度;再者,根據料兩側車道線 之車道寬度紅f彡像車道線寬度與實際車道線寬度之比值,求 成像在影像畫面上的寬度; 該偏離估算處理程賴包括頂步驟1算處理步驟··將車道雜型 利用二次曲線擬合方程式估算出實際行車路徑執跡、行進路線斜率、以及 車道曲率’湘彳T車路錄跡以及行祕祕轉算ώ車輛橫向位移量; 判斷預警倾:料輛橫向⑽量鱗道線難進行輯,若鱗靠近車 道線’則發出訊號警示; 本創作之車道猶識步驟係可針對祕畫面龍上半部進行亮度分 析,以判別日、制,並可切換價測模式提高辨識率;而影像畫面資料下 半部分為兩部份,叫同解析度進行侧,可Α幅提昇處理速度且不失债 測辨識的醉度;而該車道職步_可透過二次鱗擬合方程式,求取 車道線與消失點,並計算車輛橫向位移量及道路曲率,即使車道線受其他 車輛遮蔽阻擋,仍可利用習知之直線擬合方式進行辨識推斷;本發明亦透 過R〇I以及消失點位置偵測,進行多道反向推算,以避免誤判及不穩定 的匱况發生;且本案亦可辨識更新道路寬度,不需要事先求取車輛或攝影 200922816 機與路面之夾肖 >此係可比對車道線與實際行車路徑,當行車路徑逼近 車道線而駕驶者未有任何反應,如煞車、方向燈等反應,該信號輸出單元 係發送警訊娜駕駛者;反之,駕駛者若有反應,靡續進行偵測分析路 兄再者本發明係藉由一次曲線擬合方程式,求取車道曲率,當車速過 问且車道曲率過大時’該錢輸料元亦發岭訊,娜駕驶者放慢車速, 或進一步控制車速或煞車; 因此本發明可說是-_當具有實雜及進步性之創作,相當值得產 業界來推廣,並公諸於社會大眾。 【實施方式】 本發明係有關—種「車輛偏移之檢知方法絲置」,該車麵移之檢知 方法其主要係包括影像辨識處理程序以及偏離估算處理程序,請參照第一 圖所示,其中: 該影像辨鱗理料係·二次曲雜合方程式建構車道線模型,以 便於進行各類_之用,若實際車道較其鱗輛遮蔽,仍可彻習知之 直線擬合方式求得車道線’自於絲擬合方絲屬胃知,*贿述奸 像辨識處理程序係包括下列步驟: ϋ擷取車輛前方道路影像 影像擷取步驟:利用攝像單元安裝於車輛, 晝面資料; 以 車道線辨識步驟:將攝像單元所擷取之影像畫面資料區分為上、1 及下半部,其中: 干區間 如第一圖至第四圖所示,該影像晝面資料之上半部係分隔為— 200922816 進行日夜間判斷,並利用下列公式判斷是否為白天或夜晚,以便於切換道 路線偵測模式: ΣΣ^<^ ΣΣ^<&- ~—~<Th^ => Night ——>1bW=>Day x y gary :每個影像像素(Pixei)的灰階亮度值、thdark :暗的亮度值之閥值 (threshlod)、Thnight :判定晚上的亮度值之閥值、:判定白天的亮度 值之閥值;當灰階亮度值小於Thnight時,則判定屬於夜晚,反之則判定屬於 白天; 當灰階壳度值大於处吻時,則判定屬於白天的亮度值;反之,若灰階 亮度值小於Thnight,則判定為夜晚的亮度值; 該影像畫面資料之下半部係再分為上下二部份,影像畫面資料距離車 輛較近端之部份(即較下雜)雜驗鑛減,距離車輛較遠端之部 份(即較上部分)係維持原解析度; 本發明之車道線辨識步驟係利用一辨識流程檢測車道線是否無誤,該 辨識流程係至少包括三種車道線辨識方式: 尚灰階值辨識:利用車道線相較於路面具有較大灰階值,區別出車道 線與路面之差別; 車道線邊緣特性辨識:利用車道線與路面交接邊緣之邊緣特性,計算 標記出可能之車道線範圍; 車道線寬度辨識:彻兩側道輯之車道寬度乘上影像畫面資料車道 線之見度與實際車道線寬度之比值,求得車道線寬度祕於影像晝面資料 200922816 之寬度,藉此得知車道線寬度之判定區間; 藉此’攝像單元賴取之影像畫面資料必須同時符合上述 式’才得關找正软車魏; 職方 如第五圖’該車道線辨識步驟係將影像畫面資料下半部係由下而上 劃分為若干等分區間,並由下而上進行偵測修正判斷,如第六圖所示: 起始點搜尋流程:由最下方之區間搜尋兩側車道線之起始點作為起 點’若無法尋獲’麵續往下—區間持續尋找,直到找到起點為止; 偵測車道線流程:在目前區上述辨識流簡斷出車道線, 並利用二次擬合鱗方程式進行車道線趨勢雜,連接所有片段之車道線 做出二次曲線,該二次擬合曲線方程式之公式係為户、 /分另|1㈣際™之縱轴及橫軸為參數分別為 25㈣並完成初估之 車道線後,繼續進行軌道修正步驟,上列係數初始值會因應用不同而做調 整; 轨道修正流程:參考第六圖所示,將單—區間分割為若干列(r〇w), 並逐列進行R〇 !(邮Gn Qf inte赠,錢械圍_細靡, 修正車道線膽雜實際路徑,直到實際路軸車道線趨勢雜一致為 止,並配合車賴合航妨決策,若車雜錄況正常制續進行下一 區間之偵測車道線流程,R〇I偵測之公式如下.200922816 IX. Description of the invention: [Technical field of invention] The present invention relates to a method and device for detecting vehicle displacement, in detail, a method for judging a car secret using an image, and a quadratic rotation fitting equation The wire takes the amount of movement and the curvature of the lane to judge whether the driving is normal and can provide warning effects. [Prior Art] With the widespread use of car_humanization in human society, the gap between localities has been shortened and gradually become a social form of globalization-body, but the so-called-------in the past few years, the incidents of traffic accidents have emerged in an endless stream. Laijia and each A vehicle decoration seeks to apply New Technology to the vehicle, private: the safety of the vehicle, and reduce the incidence of traffic accidents; the occurrence of traffic accidents, often human factors account for the majority, fatigue and points God is often the main cause of accidents. Although relevant institutions at home and abroad are devoted to sleep detection, brain waves and other related research to remind drivers, but the effect is not satisfactory, coupled with the line reaction time and overall cost, can not be fully popular; In the past, the use of ultrasonic radar or light reflection and other physical means to achieve the body protection warning effect; such as the Republic of China Patent Application No. 09221746 0 "can avoid the driver's departure from the lane warning device", mainly in A vehicle warning is installed on each side of the vehicle; the vehicle warning device is provided with a voltage power supply, and a A light quantity sensor, a signal amplifier, a voltage comparison amplifier and a warning display are combined; wherein the light quantity sensor senses the amount of light reflected by the road dividing line, and the amount of light reflected by the road dividing line is greater than the road is not separated The line portion causes the light quantity sensor to generate a current value that is not the same as 200922816 by the difference in the amount of reflected light. The current value is modulated by the signal amplifier, and the current value is converted into a suitable voltage difference, and then the voltage comparison amplifier performs The voltage is relatively amplified, and a voltage output signal is obtained, and the voltage output signal is sent to the warning display ^, and the warning device is used to remind the driver to remind the driver that the vehicle being driven has deviated from the driving lane. In recent years, due to the flexibility of software writing, the convenience of c cD and CM0 s photography frames, the diversification of data collected, and the low cost, the new technology development direction is provided, so it is matched with CCD and CMO S. The high-efficiency image processing technology is developing, in addition to the reversing image, the image processing secret can hide other widely Anti-drop assistance, such as detecting vehicle road information, estimating the current driving lane of the vehicle, etc.; such as the "Way Offset Warning Device and Method" of the Republic of China Patent Application No. 0 9 213 4 0 5 The invention includes a one-way video information capturing unit, a lane offset side unit, and a __alarm unit; and the method for determining the lane offset includes image input, image analysis processing, search setting, lane offset insult, and The step of shifting the warning lane from the level or off, the automatic fine vehicle improperly shifts the lane situation' and gives the driver a warning to turn the traffic safely, and further includes an automatic judgment mechanism to determine the temporary _ warning device to avoid frequent error action. Increasing the trouble of driving people; however, this kind of technology is based on road markings and driving routes to confirm whether the vehicle is offset by positive f-orbit; but this method of comparison is not nothing--missing, there is still comparison The difference between the two causes the wrong judgment. In addition, the actual road markings and the driving route have not been estimated and calculated by the trajectory, which is expected to produce a real f deviation, and send a wrong ride, which is too late. The warning is not conducive to the actual use; therefore, in view of the problems existing in the above-mentioned conventional technologies, how to develop a more ideal and realistic structure of 200922816, the use of consumers is eagerly awaited, and the relevant industry must Efforts to develop breakthrough goals and directions. In view of this, the inventors have been engaged in the manufacturing development and design experience of related products for many years, and have made a practical and practical creation after the detailed design and careful evaluation of the above-mentioned objectives. The purpose record is purely for the detection method of the offset-based offset, which can improve the conventional technique and is limited to the system _ short, causing the reaction _ deficiency, and even due to the poor comparison method, the misjudgment and the sending of the error signal are not conducive to Actual use is missing. In order to achieve the foregoing object, the device of the present invention includes an imaging unit 'operation platform, a signal output unit, and the method includes an image recognition processing program and a deviation estimation processing program, wherein: the image recognition processing conversion system includes the following steps. . Shadow taking step: the camera unit takes the road image in front of the vehicle; the lane line identification step: the image picture data = the upper half and the lower half, like the upper half of the side data for day and night judgment The lower half of the image data is further divided into upper and lower parts. The image data is shifted from the nearer part of the vehicle to a lower resolution, and the farther part of the vehicle maintains the original resolution. Identification process, including three lane line identification methods: high gray level value identification, lane line edge characteristic identification, lane line width identification, must meet the above three detailed methods, can be judged as the correct lane line; The half system is divided into several equal divisions and the following processes are carried out from bottom to top: starting point search process: searching for the starting point of the lane lines on both sides as the starting point; 200922816 side lane line flow: in the current interval Use the above identification process to determine the lane line, and use the quadratic fitting curve equation to predict the lane line trend, and connect the lane lines of all the segments to make two Curve; financial corrections: _ R 0 1 side correction lane _ difficult to actually static. So heavy money on the money process, with ____, money car Wei prediction model constantly correct the fit of the actual lane line trajectory; The position of the vanishing point is determined by several images and the position of the vanishing point is Hejin or _, and the accuracy of the shape is added. Furthermore, according to the ratio of the lane width of the lane line on both sides of the material to the width of the actual lane line, The width of the image on the image frame; the deviation estimation process includes the top step 1 calculation process. · The lane miscellaneous equation uses the quadratic curve fitting equation to estimate the actual driving path detour, the travel route slope, and the lane curvature' Xianglu T car road recording and line secrets ώ vehicle lateral displacement; Judging early warning: material horizontal (10) scale line is difficult to carry out, if the scale is close to the lane line, then signal warning; The step of judging can perform brightness analysis on the upper part of the secret picture dragon to distinguish the day and system, and can switch the price measurement mode to improve the recognition rate. The lower part of the image picture data is two parts, called the same resolution. On the line side, the processing speed can be increased without losing the drunkness of the debt measurement identification; and the lane step _ can be used to fit the equation by the second scale, find the lane line and the vanishing point, and calculate the lateral displacement of the vehicle and the road. Curvature, even if the lane line is blocked by other vehicles, the conventional line fitting method can be used for identification and inference; the present invention also performs multi-channel backward estimation through R〇I and vanishing point position detection to avoid misjudgment and no A stable situation occurs; and the case can also identify the updated road width without the need to obtain the vehicle or the photography 200922816 machine and the road surface. This system can compare the lane line with the actual driving path, when the driving path approaches the lane line. The driver did not react, such as braking, direction lights, etc., the signal output unit sent the police driver; otherwise, if the driver responded, the follow-up analysis continued. A curve fitting equation is used to obtain the curvature of the lane. When the speed of the vehicle is too high and the curvature of the lane is too large, the money is also sent to Lingxun, and the driver slows down the speed, or further Controlling the speed of the car or braking; therefore, the invention can be said to be - when it is a complex and progressive creation, it is worthy of promotion by the industry and publicized to the public. [Embodiment] The present invention relates to a "vehicle shift detection method wire arrangement", and the vehicle surface movement detection method mainly includes an image recognition processing program and a deviation estimation processing program, please refer to the first figure. Show, among them: The image discriminant system and the quadratic curve hybrid equation construct the lane line model, so as to facilitate the use of various types of _, if the actual lane is shaded by its scale, it can still be familiar with the straight line fitting method. The lane line is obtained from the silk fitting of the square wire. The bribe description method includes the following steps: capturing the image of the road ahead of the vehicle. Steps: using the camera unit to mount the vehicle, the surface Data; lane line identification step: the image data captured by the camera unit is divided into upper, first and lower halves, wherein: the dry interval is as shown in the first to fourth figures, and the image is above the data The half system is divided into - 200922816 for day and night judgment, and use the following formula to determine whether it is day or night, in order to switch the road line detection mode: ΣΣ^<^ ΣΣ^<&- ~-~<Th ^ => ; Night ——>1bW=>Day xy gary : Grayscale brightness value of each image pixel (Pixei), threshold: Threshold of dark brightness value (threshlod), Thnight: Threshold of determining brightness value at night , determining the threshold value of the brightness value during the day; when the gray level brightness value is less than Thnight, it is determined to belong to the night, and vice versa, it is determined to belong to the day; when the gray scale shell value is greater than the kiss, the brightness value corresponding to the daytime is determined; On the other hand, if the grayscale brightness value is less than Thnight, it is determined as the nighttime brightness value; the lower half of the image frame data is further divided into upper and lower parts, and the image picture data is closer to the near part of the vehicle (ie, lower) Miscellaneous), the portion of the vehicle that is farther away (ie, the upper portion) maintains the original resolution; the lane recognition step of the present invention uses an identification process to detect whether the lane line is correct, and the identification process is at least Including three lane line identification methods: The gray scale value identification: the lane line has a larger gray scale value than the road surface, distinguishing the difference between the lane line and the road surface; lane line edge characteristic identification: using the lane line and the road surface The edge characteristics of the intersection edge are calculated, and the possible lane line range is calculated. The lane line width identification: the lane width of the two sides is multiplied by the ratio of the visibility of the image frame data lane line to the actual lane line width, and the lane line is obtained. The width is secreted by the width of the image surface data 200922816, so as to know the judgment interval of the lane line width; thereby the image of the image taken by the camera unit must meet the above formula at the same time to get the right soft car Wei; As shown in the fifth figure, the lane recognition step divides the lower part of the image frame data from bottom to top into a number of equal partitions, and performs detection and correction judgment from bottom to top, as shown in the sixth figure: Point search process: search for the starting point of the lane line on both sides from the bottom section as the starting point 'If you can't find the 'continuation down' section, continue to search until the starting point is found; Detect lane line flow: in the current area Identify the flow to break the lane line, and use the second-fit scale equation to make the lane line trend miscellaneous, and connect the lane lines of all the segments to make a quadratic curve. The formula of the program is for the household, / separate | 1 (four) of the vertical axis of the TM and the horizontal axis are 25 (four) and complete the initial evaluation of the lane line, continue the orbit correction step, the initial value of the upper coefficient will be different due to the application Making adjustments; Track correction process: Referring to the sixth figure, the single-interval is divided into several columns (r〇w), and R〇 is performed column by column! (Gn Qf inte gift, money machinery circumference _ fine, correction The actual path of the lane line is ambiguous until the actual road axis lane line trend is consistent, and the decision is made in conjunction with the vehicle navigation. If the vehicle recording condition is normal, the detection process of the next section is continued. The formula is as follows.
= [«,_! - λη · Mark j, μ ROId - [u^-λα Mark^u + λη -A/arky] + λά - Markj ] 前一列車道線的橫座標;4:前一 列沒有偵測到車道線的參數; 11 200922816 “刖歹J有制到車道線的參數:杨ι目前正在處理的列之影像平 面上的車道線寬度。 如此重複it行上述雜’赠截車道線删觀,並使車道線綱 模型不斷修正符合魏車道線軌跡; 由於車逼辨識步驟所獲得之車道線預測模型兩側車道線係於遠端交會 形成消失點,利用連續數張影像畫面資料狀消失點位置是否鄰近或相 同若/肖失點位置差異甚大則代表誤判或不穩定,藉此增加判定準域度; 又,如第七圖所示’在實際搜尋車道線時,會因車道線為虛線或標示 不月而ie成;ί〗用#像搜尋出來的點不足,無法正確擬合出來的車道線; 因此利用⑹像4面資料與實際情況之車道線寬度_,換算兩側車道線之 車道寬度乘场像車道線寬度與實際車道線寬度之比值,求得車道寬度成 像在5V像畫面上的見度,當做影像補點的依據,係當左邊有搜尋到點右邊 無時’會將左邊的點加上影像車道寬度,其位置為右邊車道線要補點的位 置’藉此得知車道線寬度之判定關以更新二次曲線車道線模型:= [«,_! - λη · Mark j, μ ROId - [u^-λα Mark^u + λη -A/arky] + λά - Markj ] The abscissa of the previous train line; 4: The previous column has no detect Measured the parameters of the lane line; 11 200922816 “刖歹J has the parameters of the lane line: the width of the lane line on the image plane of the column currently being processed by Yang. This repeats the line of the above-mentioned miscellaneous’ And the lane line model is continuously corrected to conform to the Wei lane line trajectory; the lane line of the lane line prediction model obtained by the vehicle forcing identification step is formed at the far-end intersection to form a vanishing point, and a plurality of successive image frames are used to disappear. Whether the position is adjacent or the same, if the position of the missing point is very large, it means misjudgment or instability, thereby increasing the judgment criterion. Also, as shown in the seventh figure, when the lane line is actually searched, the lane line is dotted. Or mark the month without the moon; ί〗 Use ## search for the lack of points, can not correctly fit the lane line; therefore use (6) like the 4-sided data and the actual situation of the lane line width _, convert the two lane lines Lane width multiplier like lane line width and The ratio of the actual lane line width is obtained, and the visibility of the lane width image on the 5V image is obtained. As the basis of the image complement point, when there is no search right point on the left side, the left point is added to the image lane width. Its position is the position of the right lane line to be complemented by 'by knowing the determination of the lane line width to update the quadratic lane line model:
Mark— Mark_,Mark—Mark_,
Road imagpRoad imagp
Road =>Road realRoad =>Road real
Road realRoad real
Mariw Mark」Mariw Mark
RoacU :標準兩側道路線之車道寬度,標準約為3 7 〇公分、㈣⑹:真 實道路線t ’鮮為1 5公分、RQadiliage ··影像晝面資料賴道路線之車道 寬度、Marki«age:影像畫面資料道路線寬。 如第八圖所示,當車道線模型完成後,係將已獲得之車道線模型輸入 偏離估算處理程序,並利用二次曲線擬合方程式估算出車道偏移狀態,該 偏離估算處理程序係包括下列步驟: 12 200922816 運算處理步驟:將車道賴型·二次曲線擬合方程式,估算出目前 車輛橫向位移量、車道斜率以及車道曲率; 該二次曲線擬合方程式之公式係為㈣·^分別為實 際平面朗之縱滅難,丨21、㈣為參數由上述車道_型求得. 藉由車道線可縛求得車道斜率,該車道斜率之公式係為:RoacU: The lane width of the standard road lines on both sides, the standard is about 3 7 cm, (4) (6): the real road line t 'fresh is 15 cm, RQadiliage · · image data on the road line width, Marki «age: Image line data road line width. As shown in the eighth figure, when the lane line model is completed, the obtained lane line model input is deviated from the estimation processing program, and the lane offset state is estimated by using the quadratic curve fitting equation, and the deviation estimation processing program includes The following steps: 12 200922816 Operation processing steps: The lane-dependent and quadratic curve fitting equations are used to estimate the current vehicle lateral displacement, lane slope and lane curvature; the formula of the quadratic curve fitting equation is (4)·^ respectively For the actual plane, it is difficult to find out, 丨21, (4) is the parameter obtained from the above lane _ type. The lane slope can be obtained by the lane line, and the formula of the lane slope is:
eL-2-k'X^m I 如第九圖所示,利用上述車遣料 斜羊,可以進一步推算車輛橫向位移量 △’其公式係為:△ = }广Ζχε,λ . 他々 £ 雜躺轉量,Λ :參考點之車輛 偏移量、Ζ :職距離、··行進路線斜率。 2-k 又’根據該二㈣線擬合方輯所轉行進路絲跡,村推算求得 車道曲率,該車道曲率之公式係為 {^(2-k.x + mf)n ' 因此,如第十圖所干 -(—+yL-Lx£L) 入〜 可進而推算車輛距車道邊線距離eL-2-k'X^m I As shown in the ninth figure, the vehicle lateral displacement △' can be further calculated by using the above-mentioned vehicle slanting sheep. The formula is: △ = } Ζχ ε, λ. Miscellaneous entanglement, Λ: Vehicle offset at the reference point, Ζ: duty distance, · · travel route slope. 2-k and 'according to the two (four) line fitting square track, the village road calculated to obtain the curvature of the lane, the formula of the curvature of the lane is {^(2-kx + mf)n ', therefore, as the tenth Figure dry-(-+yL-Lx£L) into ~ can further calculate the distance from the vehicle to the lane edge
Ay:Ay:
Ay = (|_八+/^)其•車道寬度(係由影像平面中某-列的影 像車道寬度透過座標轉換 得)、6K :車輛寬度、a : 路線斜率; '來其車道在影像平面的寬度由車道線模裂求 /考點之車柄偏移量、Z :預視距離、〜·行進 若車輛行駛在直、_騎,料,化如下: V 、 /、中\ ·車道寬声、A ·由* ί 了一 度〜·車輛寬度、h ··參考點之車韩 偏移量。 13 200922816 根據二次曲線擬合方程式求得之車輛橫向位移量、車道斜率以及車道 曲率,係可進入判斷預警步驟; 判斷預警步驟:料輛勤位移量鱗道、雜型進行轉崎,依昭 2 0 0 7 1 s 0 (國際標準)17 3 6 _,小客車於車道線外〇· 尺定義為最後警祕,卡車、大客車於車道料!公尺定義為最後警戒線, 而車道線㈣外财早婦极,並參考第八誠第觸,钟車路線偏 離,而靠近早《戒線或車道線時,則自紐㈣號警示;Ay = (|_eight +/^) • Lane width (converted from the image lane width of a certain column in the image plane through coordinates), 6K: vehicle width, a: route slope; 'to its lane in the image plane The width of the lane is determined by the lane line cracking/test point offset, Z: the preview distance, ~· travel if the vehicle is traveling in straight, _ riding, material, as follows: V, /, medium \ · lane wide sound , A · by * ί 1 degree ~ · vehicle width, h · · reference point car Han offset. 13 200922816 According to the quadratic curve fitting equation, the lateral displacement, lane slope and lane curvature of the vehicle can be entered into the judgment and warning step. Judgment and warning steps: the divergence of the vehicle and the miscellaneous type are carried out, 2 0 0 7 1 s 0 (International Standard) 17 3 6 _, the passenger car is outside the lane line. The ruler is defined as the last police. Trucks and buses are in the lane! The meter is defined as the last warning line, while the lane line (4) is outside the front and the front, and refers to the eighth Chengdi touch, the clock route is deviated, and when it is close to the early "cancellation or lane line, it is warning from the New (4);
14 200922816 功效。 本發明之制係還包括-種上述方法之車輛偏移檢知裝置,其主 要係包括-攝像單元、-運算平台、—信號輪出單元,其中: 當車輛啟動且車速到達-定數值時,本發明之車輛偏移檢知裝置係啟 動’該攝像單猶為-裝設於車輛内之c c D或CM〇 s攝像裝置,用以 擷取車輛射雜f彡像資訊’並騎練之魏雜魏傳輕運算平台; 該運算平自雜攝料元所娜之觀鱗魏,且車輛之車速、方 向燈及剎繼酿嫩聰㈣⑽蝴_二欠曲線擬合 林式辨識出真實道路線軌跡,藉由消失點之位置等方式檢測道路線軌跡 疋否正確’並再二次曲線擬合方程式判斷行車路徑軌跡,判斷目前車 麵及道路之姆關係,若行車路徑欲齡車舰、或車速過細彎道曲率 過大時,制淑之麵細,且額輯煞車裝置並树驗號至運算 平台’則該運算平台係發送信號至信號輸出單^反之,若達到預設之警 戒範圍’但方向燈或煞車裝置作祕發送錄至運算平台,則命令攝像單 元重新擷取影像畫面資料,重新進行判斷; 該信號輸料元縣可提供統、絲像、或音歡裝置,當運算平 台發送信紅親輸轉元’贿簡料減细錄 聲音達到提示駕敬者警訊之功效; 變化或 則藉由上述結構,當車輛行歇出現異狀,未位於車道線内、或車速過 卜而#道曲麵A ’ *易肇生意外時’本發明之車輛偏移檢知裝置係發出 警訊提示駕駛者注意,達到預警之功效者。 15 200922816 由上所述者僅為用以解釋本發明之較佳實施例,並非企圖具以對發明 作做任何形式上之關’是以,凡有在相同之發曝神下所做有關發明作 之任何修飾或變更者,為其他可據以實施之態樣且具有相同效果者,皆仍 應包括在本發明意圖保護之範疇内。 综上㈣’本發明之「車輛偏移之檢知㈣絲置」,於結構設計及使 之創2,符合實祕,謂财,錢砰前所未有 更具功效^所哗具有「新概」應無疑慮,又本發明可較之習用結構 專利申,之:進’因此亦具有「進步性」,其^全符合我國專利法有關發明 早日賜予本宰Γ故’麦依法向釣局提出發明專利申請,_狗局能 案專利權,至感德便。 16 200922816 【圖式簡單說明】 第一圖係本發明車輛偏移檢知方法流程示意圖。 第二圖至第四圖係本發明日夜判斷辨識示意圖。 第五圖係本發明車道線劃分區間及列狀態示意圖。 第六圖係本發明車道線辨識步驟流程示意圖。 第七圖係本發明利用座標轉換求得車道線寬度示意圖。 第八圖係本發明偏離估算處理程序實際車輛狀態示意圖。 第九圖係本發明車輛距車道邊線距離計算示意圖。 第十圖係本發明車輛與車道線平行時橫向位移量計算示意圖。 【主要元件符號說明】 無 1714 200922816 Efficacy. The system of the present invention further includes a vehicle offset detecting device of the above method, which mainly includes an - camera unit, a computing platform, and a signal wheeling unit, wherein: when the vehicle starts and the vehicle speed reaches a fixed value, The vehicle offset detecting device of the present invention activates the cc D or CM 〇 camera device installed in the vehicle, and is used to capture the vehicle 杂 彡 资讯 资讯Wei Wei Chuan light computing platform; the operation of the self-contained material Yuan Naozhi Guan Wei Wei, and the vehicle speed, direction lights and brakes after the production of Nen Cong (four) (10) butterfly _ two under-curve fitting forest type to identify the real road line The trajectory, whether the road line trajectory is correct by means of the position of the vanishing point, etc., and the quadratic curve fitting equation is used to determine the trajectory of the driving path, and the relationship between the current vehicle surface and the road is determined, and if the driving path is an age-old vehicle, or When the speed of the car is too large, the curvature of the curve is too large, and the surface of the system is fine, and the amount of the brake device is set to the calculation platform. The computing platform sends a signal to the signal output unit. Otherwise, if the preset warning range is reached, Direction light or 煞The device secretly records and records to the computing platform, and then commands the camera unit to retrieve the image image data again and re-determines; the signal transmission element county can provide a unified, silk image, or sound device, when the computing platform sends a letter to the red The transfer of the yuan's bribes to reduce the sound of the sound reaches the effect of warning the driver's warning; the change or the above structure, when the vehicle breaks, the vehicle is not in the lane line, or the speed is over. #道曲When the surface A ' * is prone to accidents, the vehicle displacement detecting device of the present invention issues a warning to the driver to pay attention to the effect of the warning. 15 200922816 The above description is only for explaining the preferred embodiment of the present invention, and is not intended to be in any form to the invention, so that all related inventions are made under the same exposure. Any modifications or alterations made by others that have been implemented and have the same effect are still included in the scope of the invention. In summary, (4) 'The detection of vehicle offset (4) silk set in the invention, in the structural design and make it 2, in line with the real secret, that is, the money, the money has never been more effective ^ has a "new" should Undoubtedly, the invention can be compared with the conventional structure patent application: it is also "progressive", and it is in line with the invention of the patent law of China, and it is given to the company as soon as possible. Application, _ dog office can patent the patent, to the sense of virtue. 16 200922816 [Simplified description of the drawings] The first figure is a schematic flow chart of the vehicle offset detection method of the present invention. The second to fourth figures are schematic diagrams of the day and night judgment identification of the present invention. The fifth figure is a schematic diagram of the lane division section and the column state of the present invention. The sixth figure is a schematic flow chart of the lane marking step of the present invention. The seventh figure is a schematic diagram of the width of the lane line obtained by the coordinate conversion of the present invention. The eighth figure is a schematic diagram of the actual vehicle state of the deviation estimation processing program of the present invention. The ninth figure is a schematic diagram of the distance calculation of the vehicle from the lane side line of the present invention. The tenth figure is a schematic diagram of the calculation of the lateral displacement amount when the vehicle of the present invention is parallel to the lane line. [Main component symbol description] None 17
Claims (1)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW096145498A TW200922816A (en) | 2007-11-30 | 2007-11-30 | Method and device for detecting the lane deviation of vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW096145498A TW200922816A (en) | 2007-11-30 | 2007-11-30 | Method and device for detecting the lane deviation of vehicle |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TW200922816A true TW200922816A (en) | 2009-06-01 |
| TWI337144B TWI337144B (en) | 2011-02-11 |
Family
ID=44728339
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW096145498A TW200922816A (en) | 2007-11-30 | 2007-11-30 | Method and device for detecting the lane deviation of vehicle |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TW200922816A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI408339B (en) * | 2010-03-22 | 2013-09-11 | Inst Information Industry | Real-time augmented reality device, real-time augmented reality methode and computer program product thereof |
| CN103885573A (en) * | 2012-12-19 | 2014-06-25 | 财团法人车辆研究测试中心 | Automatic correction method and system for vehicle display system |
| US8902053B2 (en) | 2011-11-16 | 2014-12-02 | Industrial Technology Research Institute | Method and system for lane departure warning |
| TWI549100B (en) * | 2012-08-03 | 2016-09-11 | 能晶科技股份有限公司 | Driver detection apparatus and method thereof |
| TWI621073B (en) * | 2016-12-08 | 2018-04-11 | Road lane detection system and method thereof | |
| TWI638343B (en) * | 2015-12-22 | 2018-10-11 | 豪威科技股份有限公司 | Lane detection system and method |
| US10102435B2 (en) | 2016-08-10 | 2018-10-16 | Omnivision Technologies, Inc. | Lane departure warning system and associated methods |
| CN111284487A (en) * | 2018-12-07 | 2020-06-16 | 星克跃尔株式会社 | Lane line display method and electronic device for performing the same |
| WO2023049550A1 (en) * | 2021-09-21 | 2023-03-30 | Qualcomm Incorporated | High resolution camera system for automotive vehicles |
| TWI798021B (en) * | 2022-03-10 | 2023-04-01 | 台灣智慧駕駛股份有限公司 | Method and system for measuring vehicle deviation |
| CN116229723A (en) * | 2023-05-06 | 2023-06-06 | 山东纵云信息技术有限公司 | An intelligent traffic information data analysis management system |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI498831B (en) * | 2013-10-15 | 2015-09-01 | Nat Univ Chung Hsing | System of adaptive variable shape region of interest determination and its control method |
| US9091558B2 (en) | 2013-12-23 | 2015-07-28 | Automotive Research & Testing Center | Autonomous driver assistance system and autonomous driving method thereof |
| TWI609807B (en) * | 2016-05-17 | 2018-01-01 | 緯創資通股份有限公司 | Image evaluation method and electronic apparatus thereof |
| TWI611961B (en) * | 2016-06-01 | 2018-01-21 | 緯創資通股份有限公司 | Device, method, and computer-readable medium for analyzing lane line image |
-
2007
- 2007-11-30 TW TW096145498A patent/TW200922816A/en unknown
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI408339B (en) * | 2010-03-22 | 2013-09-11 | Inst Information Industry | Real-time augmented reality device, real-time augmented reality methode and computer program product thereof |
| US8902053B2 (en) | 2011-11-16 | 2014-12-02 | Industrial Technology Research Institute | Method and system for lane departure warning |
| TWI549100B (en) * | 2012-08-03 | 2016-09-11 | 能晶科技股份有限公司 | Driver detection apparatus and method thereof |
| CN103885573A (en) * | 2012-12-19 | 2014-06-25 | 财团法人车辆研究测试中心 | Automatic correction method and system for vehicle display system |
| CN103885573B (en) * | 2012-12-19 | 2017-03-01 | 财团法人车辆研究测试中心 | Automatic correction method and system for vehicle display system |
| TWI638343B (en) * | 2015-12-22 | 2018-10-11 | 豪威科技股份有限公司 | Lane detection system and method |
| US10102434B2 (en) | 2015-12-22 | 2018-10-16 | Omnivision Technologies, Inc. | Lane detection system and method |
| US10102435B2 (en) | 2016-08-10 | 2018-10-16 | Omnivision Technologies, Inc. | Lane departure warning system and associated methods |
| TWI621073B (en) * | 2016-12-08 | 2018-04-11 | Road lane detection system and method thereof | |
| CN111284487A (en) * | 2018-12-07 | 2020-06-16 | 星克跃尔株式会社 | Lane line display method and electronic device for performing the same |
| CN111284487B (en) * | 2018-12-07 | 2024-01-23 | 星克跃尔株式会社 | Lane line display method and electronic device for executing same |
| WO2023049550A1 (en) * | 2021-09-21 | 2023-03-30 | Qualcomm Incorporated | High resolution camera system for automotive vehicles |
| TWI798021B (en) * | 2022-03-10 | 2023-04-01 | 台灣智慧駕駛股份有限公司 | Method and system for measuring vehicle deviation |
| CN116229723A (en) * | 2023-05-06 | 2023-06-06 | 山东纵云信息技术有限公司 | An intelligent traffic information data analysis management system |
| CN116229723B (en) * | 2023-05-06 | 2023-08-11 | 山东纵云信息技术有限公司 | An Intelligent Traffic Information Data Analysis and Management System |
Also Published As
| Publication number | Publication date |
|---|---|
| TWI337144B (en) | 2011-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TW200922816A (en) | Method and device for detecting the lane deviation of vehicle | |
| CN101470801B (en) | Vehicle offset detection method | |
| US12106574B2 (en) | Image segmentation | |
| CN109017780B (en) | Intelligent driving control method for vehicle | |
| CN103760569B (en) | A kind of drivable region detection method based on laser radar | |
| CN103985250B (en) | The holographic road traffic state vision inspection apparatus of lightweight | |
| CN102270301B (en) | Method for detecting unstructured road boundary by combining support vector machine (SVM) and laser radar | |
| US10885358B2 (en) | Method for detecting traffic signs | |
| EP2927081B1 (en) | Acceleration suppression device for vehicle, and acceleration suppression method for vehicle | |
| CN102336163B (en) | Vehicle yaw detection device | |
| CN102622886B (en) | Video-based method for detecting violation lane-changing incident of vehicle | |
| WO2018171464A1 (en) | Method, apparatus and system for planning vehicle speed according to navigation path | |
| CN103236191B (en) | Video-based safety precaution method for vehicle merging from highway ramp | |
| CN104608766A (en) | Automatic parking method and system used for intelligent vehicle through parking memory stick | |
| CN102997926B (en) | A kind of method for obtaining navigation data | |
| CN114715168A (en) | Vehicle yaw early warning method and system under road marking missing environment | |
| CN110472468A (en) | Monocular vision odometer: the speed and yaw-rate of the vehicle from rear view camera | |
| CN106601023B (en) | Indoor parking lot vehicle guiding and positioning system and method | |
| CN107193011A (en) | A kind of method for being used to quickly calculate car speed in automatic driving car area-of-interest | |
| CN102324183A (en) | Vehicle detection and grasp shoot method based on compound virtual coil | |
| CN105480227B (en) | A method based on fusion of infrared radar and video image information in active driving technology | |
| CN110263844B (en) | A Method for Online Learning and Real-time Estimation of Pavement State | |
| CN104044594A (en) | Arithmetic device oriented toward transverse distance early warning | |
| CN104077756A (en) | Direction filtering method based on lane line confidence | |
| CN103116988A (en) | Traffic flow and vehicle type detecting method based on TOF (time of flight) camera |