TWI646001B - Lane departure warning method and system - Google Patents
Lane departure warning method and system Download PDFInfo
- Publication number
- TWI646001B TWI646001B TW106128035A TW106128035A TWI646001B TW I646001 B TWI646001 B TW I646001B TW 106128035 A TW106128035 A TW 106128035A TW 106128035 A TW106128035 A TW 106128035A TW I646001 B TWI646001 B TW I646001B
- Authority
- TW
- Taiwan
- Prior art keywords
- lane
- deviation
- continuously
- driver
- current image
- Prior art date
Links
Landscapes
- Traffic Control Systems (AREA)
- Image Processing (AREA)
Abstract
一種車道偏離警示方法，包含以下步驟：(A)一人臉位置偵測單元持續地拍攝一駕駛者之人臉以便連續獲得該駕駛者之一當前影像，並持續地根據該當前影像及該當前影像的前N張影像，持續地獲得並傳送該駕駛者的一偏離車道傾向至一處理單元；(B)該處理單元判斷是否接收到一車輛偏離資料；(C)當該處理單元判斷出接收到該車輛偏離資料時，該處理單元判斷來自該人臉位置偵測單元的該偏離車道傾向是否與該車輛偏離資料所包含的車輛偏離方向相同；及(D)當該處理單元判斷出該偏離車道傾向與該車輛偏離方向不同時，該處理單元傳送一警示訊號至一警示單元。 A lane departure warning method includes the following steps: (A) a face position detecting unit continuously captures a driver's face to continuously obtain a current image of the driver, and continuously according to the current image and the current image The first N images continuously acquire and transmit the driver's deviation from the lane to a processing unit; (B) the processing unit determines whether a vehicle deviation data is received; (C) when the processing unit determines that the reception is received When the vehicle deviates from the data, the processing unit determines whether the off-track tendency from the face position detecting unit is the same as the deviation direction of the vehicle included in the vehicle deviation data; and (D) when the processing unit determines the off-track When the direction of departure from the vehicle is different, the processing unit transmits an alert signal to an alert unit.
Description
本發明是有關於一種車道偏離警示方法，特別是指一種利用影像辨識技術判定駕駛者之駕駛情況以決定是否發出車道偏離警示的車道偏離警示方法及其系統。 The present invention relates to a lane departure warning method, and more particularly to a lane departure warning method and system for determining whether a driver's driving situation is determined by using image recognition technology to determine whether to issue a lane departure warning.
無意的車道偏離往往容易造成與對向車道的來車正面碰撞，或與同向或反向行駛之車輛發生側面擦撞。不論是正面碰撞還是側面擦撞，都將對車上乘員的生命造成嚴重威脅。為了降低交通事故中，危險駕駛與疲勞駕駛的機率，以達到安全駕駛之目的，遂有車道偏離警示系統(Lane Departure Warning System)之研發，車道偏離警示系統，是一項汽車駕駛安全輔助系統，可用於輔助駕駛行車安全。一旦駕駛者所駕駛之車輛在非蓄意情形下開始偏移車道時，車道偏離警示系統即會自動發出聲音、視覺或觸覺等警示訊號，以提醒駕駛者應即時做出應變以預防或減少事故，並減低傷亡風險。 Unintentional lane departures tend to cause frontal collisions with incoming vehicles in opposite lanes or side collisions with vehicles traveling in the same or opposite directions. Whether it is a frontal collision or a side collision, it will pose a serious threat to the life of the occupants. In order to reduce the risk of dangerous driving and fatigue driving in traffic accidents, in order to achieve safe driving, the Lane Departure Warning System is developed and the lane departure warning system is a driving assistance system for vehicles. Can be used to assist driving driving safety. Once the driver's vehicle starts to shift lanes in an unintentional situation, the lane departure warning system will automatically emit an audible, visual or tactile warning signal to remind the driver to respond immediately to prevent or reduce the accident. And reduce the risk of casualties.
以往在使用車道偏離警示系統時，車道偏離警示系統需要收集車輛的行車資訊，以根據行車資訊判定行車方向燈是否有啟動，當車道偏離警示系統判定出車輛偏離車道且行車方向燈無啟動時，車道偏離警示系統才會發出警示。否則，當駕駛者出其本意在變換車道時，即會導致車道偏離警示系統發出假警示，但是在某些情況下，相關於行車方向燈的行車資訊不易取得，例如將車道偏離警示系統加裝於行車紀錄器時，即不易取得行車資訊，又或是駕駛者在變換車道時，忘記啟動行車方向燈，皆會導致車道偏離警示系統發出假警示，此類誤報將導致駕駛降低使用車道偏離警示系統之意願，唯有可靠並準確地發出警示，才能使車道偏離警示系統在使用上更加可靠，進而提升駕駛的使用意願，故實有必要尋求一正準率較高的車道偏離警示方法。 In the past, when using the lane departure warning system, the lane departure warning system needs to collect the driving information of the vehicle to determine whether the driving direction light is activated according to the driving information. When the lane departure warning system determines that the vehicle deviates from the lane and the driving direction light does not start, The lane departure warning system will only issue a warning. Otherwise, when the driver intends to change the lane, it will cause the lane departure warning system to issue a false warning, but in some cases, the driving information related to the driving direction light is not easy to obtain, for example, the lane departure warning system is installed. When driving the recorder, it is not easy to obtain driving information, or the driver forgets to activate the driving direction light when changing lanes, which will cause the lane departure warning system to issue false warnings. Such false alarms will lead to reduced driving lane departure warning. The willingness of the system, only the reliable and accurate warning, can make the lane departure warning system more reliable in use, and thus enhance the willingness to use the driving. Therefore, it is necessary to seek a lane departure warning method with a higher accuracy.
因此，本發明的一目的，即在提供一種方便使用且正確率較高的車道偏離警示方法。 Accordingly, it is an object of the present invention to provide a lane departure warning method that is convenient to use and has a high correct rate.
於是，本發明車道偏離警示方法，藉由一處理單元及一電連接該處理單元的人臉位置偵測單元來實施，該處理單元還電連接一車道偵測單元與一警示單元，該車道偵測單元被配置來偵測一駕駛者所駕駛之車輛行駛於一車道的狀態，以便在該車輛偏離該車道時產生一包含一車輛偏離方向的車輛偏離資料，並傳送該車輛偏 離資料至該處理單元，該車道偏離警示方法包含以下步驟：(A)該人臉位置偵測單元持續地拍攝該駕駛者之人臉以便連續獲得該駕駛者之一當前影像，並持續地根據該當前影像及該當前影像的前N張影像，持續地獲得並傳送該駕駛者的一偏離車道傾向至該處理單元，N≧1；(B)該處理單元判斷是否接收到來自該車道偵測單元的該車輛偏離資料；(C)當該處理單元判斷出接收到來自該車道偵測單元的該車輛偏離資料時，該處理單元判斷來自該人臉位置偵測單元的該偏離車道傾向是否與該車輛偏離資料所包含的車輛偏離方向相同；及(D)當該處理單元判斷出該偏離車道傾向與該車輛偏離方向不同時，該處理單元傳送一警示訊號至該警示單元。 Therefore, the lane departure warning method of the present invention is implemented by a processing unit and a face position detecting unit electrically connected to the processing unit, and the processing unit is further electrically connected to a lane detecting unit and a warning unit. The measuring unit is configured to detect a state in which the driver's vehicle is driving in a lane, so as to generate a vehicle deviation data including a vehicle deviation direction when the vehicle deviates from the lane, and transmit the vehicle bias Departing the data to the processing unit, the lane departure warning method comprises the following steps: (A) the face position detecting unit continuously captures the driver's face to continuously obtain the current image of the driver, and continuously according to The current image and the first N images of the current image continuously obtain and transmit a deviation direction of the driver to the processing unit, N≧1; (B) the processing unit determines whether the lane detection is received. The vehicle of the unit deviates from the data; (C) when the processing unit determines that the vehicle deviation data from the lane detecting unit is received, the processing unit determines whether the deviation lane tendency from the face position detecting unit is The vehicle deviation direction includes the same vehicle deviation direction; and (D) when the processing unit determines that the deviation lane tendency is different from the vehicle deviation direction, the processing unit transmits a warning signal to the warning unit.
本發明的另一目的，即在提供一種方便使用且正確率較高的車道偏離警示系統。 Another object of the present invention is to provide a lane departure warning system that is convenient to use and has a high correct rate.
於是，本發明車道偏離警示系統包含一人臉位置偵測單元、一車道偵測單元、一警示單元及一處理單元。 Therefore, the lane departure warning system of the present invention comprises a face position detecting unit, a lane detecting unit, a warning unit and a processing unit.
該人臉位置偵測單元用於持續地拍攝一駕駛者之人臉以便連續獲得該駕駛者之一當前影像，並持續地根據該當前影像及該當前影像的前N張影像，持續地獲得該駕駛者的一偏離車道傾向， N≧1。 The face position detecting unit is configured to continuously capture a driver's face to continuously obtain a current image of the driver, and continuously obtain the current image and the first N images of the current image. The driver’s tendency to deviate from the lane, N≧1.
該車道偵測單元，用於持續地偵測該駕駛者所駕駛之車輛行駛於一車道的狀態，以便在該車輛偏離該車道時產生一包含一車輛偏離方向的車輛偏離資料。 The lane detecting unit is configured to continuously detect a state in which the driver's vehicle is driving in a lane, so as to generate a vehicle deviation data including a vehicle deviation direction when the vehicle deviates from the lane.
該警示單元用於發出一警示。 The alert unit is used to issue a warning.
該處理單元電連接該人臉位置偵測單元、該車道偵測單元與該警示單元，且持續地接收來自該人臉位置偵測單元的偏離車道傾向，並接收來自該車道偵測單元的車輛偏離資料。 The processing unit is electrically connected to the face position detecting unit, the lane detecting unit and the warning unit, and continuously receives the off-track tendency from the face position detecting unit, and receives the vehicle from the lane detecting unit Deviation from the information.
其中，該處理單元判斷是否接收到來自該車道偵測單元的該車輛偏離資料，當該處理單元判斷出接收到來自該車道偵測單元的該車輛偏離資料時，該處理單元判斷來自該人臉位置偵測單元的該偏離車道傾向是否與該車輛偏離資料所包含的車輛偏離方向相同，當該處理單元判斷出該偏離車道傾向與該車輛偏離方向不同時，該處理單元傳送一警示訊號至該警示單元。 The processing unit determines whether the vehicle deviation data from the lane detection unit is received. When the processing unit determines that the vehicle deviation data from the lane detection unit is received, the processing unit determines that the vehicle face is from the face. Whether the deviation direction of the position detecting unit is the same as the deviation direction of the vehicle included in the vehicle deviation data, and when the processing unit determines that the deviation lane tendency is different from the vehicle deviation direction, the processing unit transmits a warning signal to the Warning unit.
本發明的功效在於：藉由利用該駕駛者之影像來判定該偏離車道傾向與該車輛偏離資料所包含的車輛偏離方向是否同向可更加直覺且正確地判斷出該駕駛者是否為蓄意偏移車道，且自該人臉位置偵測單元獲得該駕駛者之偏離車道傾向，與自行車電腦取得行車資訊相較下，較為容易且方便。 The effect of the present invention is to determine whether the driver's deliberate deviation is more intuitively and correctly determined by using the driver's image to determine whether the off-track tendency and the deviation direction of the vehicle included in the vehicle deviation data are in the same direction. The lane, and the driver's deviation from the lane is obtained from the face position detecting unit, which is easier and more convenient than the bicycle computer to obtain driving information.
1‧‧‧人臉位置偵測單元 1‧‧‧Face position detection unit
2‧‧‧車道偵測單元 2‧‧‧ Lane Detection Unit
3‧‧‧警示單元 3‧‧‧Warning unit
4‧‧‧處理單元 4‧‧‧Processing unit
61~65‧‧‧步驟 61~65‧‧‧Steps
621~629‧‧‧子步驟 621~629‧‧‧ substeps
本發明的其他的特徵及功效，將於參照圖式的實施方式中清楚地呈現，其中：圖1是本發明車道偏離警示系統之實施例的一方塊圖；圖2是一流程圖，說明本發明車道偏離警示方法之實施例；圖3是一流程圖，說明一人臉位置偵測單元如何根據一當前影像及該當前影像的前N張影像來獲得一駕駛者之一偏離車道傾向的細部流程；及圖4是一流程圖，說明該人臉位置偵測單元如何根據一普通狀態之總數量、一意圖右偏狀態之總數量及一意圖左偏狀態之總數量來獲得該偏離車道傾向的細部流程。 Other features and advantages of the present invention will be apparent from the embodiments of the present invention. FIG. 1 is a block diagram of an embodiment of the lane departure warning system of the present invention. FIG. 2 is a flow chart illustrating the present invention. An embodiment of the invention discloses a lane departure warning method. FIG. 3 is a flowchart illustrating how a face position detecting unit obtains a detailed flow of a driver's deviation from a lane based on a current image and the first N images of the current image. And FIG. 4 is a flow chart showing how the face position detecting unit obtains the off-track tendency according to the total number of a normal state, the total number of intent right deviation states, and the total number of intention left deviation states. Detailed process.
參閱圖1，本發明車道偏離警示系統的實施例包含一人臉位置偵測單元1、一車道偵測單元2、一警示單元3、及一處理單元4。 Referring to FIG. 1 , an embodiment of the lane departure warning system of the present invention includes a face position detecting unit 1, a lane detecting unit 2, a warning unit 3, and a processing unit 4.
該人臉位置偵測單元1用於持續地獲得一駕駛者的一偏離車道傾向，並傳送該偏離車道傾向至該處理單元4。在本實施例中，該人臉位置偵測單元1包含一影像拍攝模組(圖未示)，及一電連接該影像拍攝模組的偏離車道傾向判定模組(圖未示)。該影像拍 攝模組用於拍攝該駕駛者之人臉以便連續獲得該駕駛者之一當前影像。在該偏離車道傾向判定模組接收到來自該影像拍攝模組的該當前影像後，該偏離車道傾向判定模組判定是否自該當前影像辨識出對應於該駕駛者之人臉的一眼部位置及/或一鼻子位置。當該偏離車道傾向判定模組判定出無法自該當前影像辨識出對應於該駕駛者之人臉的該眼部位置及/或該鼻子位置時，該偏離車道傾向判定模組將該當前影像標記為一普通狀態，且獲得一指示出一未主動偏離車道的偏離車道傾向，並傳送該偏離車道傾向至該處理單元4。當該偏離車道傾向判定模組判定出成功地自該當前影像辨識出對應於該駕駛者之人臉的該眼部位置及/或該鼻子位置時，該偏離車道傾向判定模組根據所辨識出之該當前影像中對應於該駕駛者之人臉的該眼部位置及/或該鼻子位置所構成的幾何形狀或遮蔽程度，獲得該駕駛者之人臉的一臉部轉向程度。該偏離車道傾向判定模組根據該當前影像所對應的臉部轉向程度，及先前所獲得之該當前影像的前N張影像所對應的臉部轉向程度，獲得該駕駛者的一偏離車道傾向。該偏離車道傾向指示出一主動朝右側偏離、一主動朝左側偏離或該未主動偏離車道等三種狀態之其中一者。在本實施例中，該人臉位置偵測單元1係配合一基於哈爾特徵訓練的階層式自適應增強(Haar-like Adaboost Cascade)演算法來辨識該當前影像中對應於該駕駛者之人臉的一眼部位置及/或一鼻子位置(圖 未示)，但不以此為限。 The face position detecting unit 1 is configured to continuously obtain a driver's deviation from the lane and transmit the off-track tendency to the processing unit 4. In this embodiment, the face position detecting unit 1 includes an image capturing module (not shown) and an off-track tendency determining module (not shown) electrically connected to the image capturing module. The image shoot The camera module is used to capture the driver's face to continuously obtain the current image of the driver. After the off-track tendency determination module receives the current image from the image capturing module, the off-track tendency determination module determines whether an eye position corresponding to the driver's face is recognized from the current image. And / or a nose position. When the off-road tendency determination module determines that the eye position and/or the nose position corresponding to the driver's face cannot be recognized from the current image, the off-track tendency determination module marks the current image It is a normal state, and a deviation from the lane indicating a non-active departure from the lane is obtained, and the deviation lane is transmitted to the processing unit 4. When the off-road tendency determination module determines that the eye position and/or the nose position corresponding to the driver's face is successfully recognized from the current image, the deviation lane tendency determination module is identified according to the The degree of turning of the face of the driver's face is obtained in the current image corresponding to the eye position of the driver's face and/or the geometric shape or degree of obscuration formed by the nose position. The off-road tendency determination module obtains a deviation tendency of the driver from the driver according to the degree of face steering corresponding to the current image and the degree of face steering corresponding to the previous N images of the current image obtained. The off-track tendency indicates one of three states, such as an active deviation to the right, an active deviation to the left, or a non-active departure from the lane. In this embodiment, the face position detecting unit 1 cooperates with a Haar-like Adaboost Cascade algorithm based on Haar feature training to identify the person corresponding to the driver in the current image. One eye position of the face and/or a nose position (figure Not shown), but not limited to this.
該車道偵測單元2用於持續地偵測該駕駛者所駕駛之車輛行駛於一車道的狀態，當該車道偵測單元2偵測出該車輛偏離該車道時產生一包含一車輛偏離方向的車輛偏離資料，並傳送該車輛偏離資料至該處理單元4。在本實施例中，該車道偵測單元2包含一用於拍攝該車輛所行駛之車道以便連續產生該車道之車道影像的影像拍攝模組(圖未示)，及一電連接該影像拍攝模組的車道偏離判定模組(圖未示)。在該車道偏離判定模組接收到來自該影像拍攝模組的車道影像後，該車道偏離判定模組利用一習知的車道線辨識技術，辨識出所接收之車道影像的車道線部分，並根據所辨識出的車道線部分判定該車輛是否偏離該車道。當該車道偏離判定模組判定出該車輛偏離該車道時，該車道偏離判定模組產生包含該車輛偏離方向的車輛偏離資料，其中該車輛偏離方向指示出一朝右側偏離、一朝左側偏離等二種狀態之其中一者，並傳送該車輛偏離資料至該處理單元4，且繼續根據所接收到的來自該影像拍攝模組的下一車道影像辨識出車道線部分，並根據所辨識出的車道線部分判定該車輛是否偏離該車道；當該車道偏離判定模組判定出該車輛無偏離該車道時，該車道偏離判定模組繼續根據所接收到的來自該影像拍攝模組的下一車道影像辨識出車道線部分，並根據所辨識出的車道線部分判定該車輛是否偏離該車道。由於本發明之特徵並不在於熟知 此技藝者所已知的該車道偏離判定模組之車道偏離判定方式，為了簡潔，在此省略了其細節。 The lane detecting unit 2 is configured to continuously detect a state in which the driver's vehicle is driving in a lane. When the lane detecting unit 2 detects that the vehicle deviates from the lane, the vehicle detecting unit 2 generates a vehicle deviation direction. The vehicle deviates from the data and transmits the vehicle deviation data to the processing unit 4. In this embodiment, the lane detecting unit 2 includes an image capturing module (not shown) for taking a lane of the vehicle to continuously generate a lane image of the lane, and electrically connecting the image capturing module. The lane departure determination module of the group (not shown). After the lane departure determination module receives the lane image from the image capturing module, the lane departure determination module uses a conventional lane line identification technology to identify the lane line portion of the received lane image, and according to the The identified lane line portion determines whether the vehicle deviates from the lane. When the lane departure determination module determines that the vehicle deviates from the lane, the lane departure determination module generates vehicle deviation data including the vehicle deviation direction, wherein the vehicle deviation direction indicates a rightward deviation, a leftward deviation, and the like. One of the two states, and transmitting the vehicle deviation data to the processing unit 4, and continuing to recognize the lane line portion according to the received next lane image from the image capturing module, and according to the identified The lane line portion determines whether the vehicle deviates from the lane; when the lane departure determination module determines that the vehicle does not deviate from the lane, the lane departure determination module continues to follow the received next lane from the image capturing module The image identifies the lane line portion and determines whether the vehicle deviates from the lane based on the identified lane line portion. Since the features of the present invention are not well known The lane departure determination mode of the lane departure determination module known to the skilled person is omitted here for brevity.
該警示單元3用於發出一警示。在本實施例中，該警示為一聲音訊號、一影像訊號或兩者的結合，但不限於此。 The warning unit 3 is used to issue a warning. In this embodiment, the alert is an audio signal, an image signal, or a combination of the two, but is not limited thereto.
該處理單元4與該人臉位置偵測單元1、該車道偵測單元2、該警示單元3電連接，且該處理單元4可持續地接收來自該人臉位置偵測單元1的偏離車道傾向，並接收來自該車道偵測單元2的車輛偏離資料。 The processing unit 4 is electrically connected to the face position detecting unit 1, the lane detecting unit 2, and the warning unit 3, and the processing unit 4 can continuously receive the off-track tendency from the face position detecting unit 1. And receiving vehicle deviation data from the lane detecting unit 2.
參閱圖1與圖2，以下將配合本發明車道偏離警示方法之實施例，來說明該車道偏離警示系統中各元件的運作細節，該車道偏離警示方法之實施例包含下列步驟。 Referring to Figures 1 and 2, the operation details of the components in the lane departure warning system will be described below in conjunction with the embodiment of the lane departure warning method of the present invention. The embodiment of the lane departure warning method includes the following steps.
在步驟61中，該車道偵測單元2透過該影像拍攝模組及該車道偏離判定模組(圖未示)持續地偵測該駕駛者所駕駛之車輛行駛於該車道的狀態，以判定該車輛是否偏離該車道。 In step 61, the lane detecting unit 2 continuously detects the state in which the driver's vehicle is driving in the lane through the image capturing module and the lane departure determining module (not shown) to determine the state. Whether the vehicle deviates from the lane.
在步驟62中，該人臉位置偵測單元1持續地獲得該當前影像，並持續地根據該當前影像及該當前影像的前N張影像獲得並傳送該駕駛者的該偏離車道傾向至該處理單元4，其中N為正整數，亦即N≧1。 In step 62, the face position detecting unit 1 continuously obtains the current image, and continuously obtains and transmits the driver's off-track tendency according to the current image and the first N images of the current image to the process. Unit 4, where N is a positive integer, that is, N≧1.
參閱圖1與圖3，值得特別說明的是，步驟62還包含以下子步驟。值得注意的是，以下將針對該人臉位置偵測單元1獲得該 當前影像後，如何根據該當前影像及該當前影像的前N張影像獲得該駕駛者的該偏離車道傾向作細部說明，由於該人臉位置偵測單元1會持續地獲得當前影像，故後續該人臉位置偵測單元1在獲得當前影像後，亦會持續地進行步驟621~626。 Referring to Figures 1 and 3, it is worthwhile to note that step 62 also includes the following sub-steps. It is worth noting that the following will be obtained for the face position detecting unit 1 After the current image, how to obtain the driver's deviation from the lane based on the current image and the first N images of the current image is described as a detailed description. Since the face position detecting unit 1 continuously obtains the current image, the subsequent image is After obtaining the current image, the face position detecting unit 1 continues to perform steps 621-626.
在步驟621中，該人臉位置偵測單元1的該影像拍攝模組獲得並傳送該當前影像至該人臉位置偵測單元1的該偏離車道傾向判定模組，當該偏離車道傾向判定模組接收到來自人臉位置偵測單元1之該影像拍攝模組的該當前影像時，該偏離車道傾向判定模組判定是否自該當前影像辨識出對應於該駕駛者之人臉的該眼部位置及/或該鼻子位置。當該偏離車道傾向判定模組判定出無法自該當前影像辨識出對應於該駕駛者之人臉的該眼部位置及/或該鼻子位置時，流程進行步驟622，當該偏離車道傾向判定模組判定出成功地自該當前影像辨識出對應於該駕駛者之人臉的該眼部位置及/或該鼻子位置時，流程進行步驟623。 In step 621, the image capturing module of the face position detecting unit 1 obtains and transmits the current image to the off-track tendency determining module of the face position detecting unit 1. When the group receives the current image from the image capturing module of the face position detecting unit 1, the deviation lane tendency determining module determines whether the eye corresponding to the driver's face is recognized from the current image. Location and / or the nose position. When the deviation lane tendency determination module determines that the eye position and/or the nose position corresponding to the driver's face cannot be recognized from the current image, the flow proceeds to step 622, when the deviation lane tendency determination mode is performed. When the group determines that the eye position and/or the nose position corresponding to the driver's face is successfully recognized from the current image, the flow proceeds to step 623.
在步驟622中，該偏離車道傾向判定模組將該當前影像標記為該普通狀態，且獲得指示出該未主動偏離車道的該偏離車道傾向，並傳送該偏離車道傾向至該處理單元4。 In step 622, the off-road tendency determination module marks the current image as the normal state, and obtains the off-track tendency indicating the non-active departure from the lane, and transmits the off-track tendency to the processing unit 4.
在步驟623中，該偏離車道傾向判定模組根據所辨識出之該當前影像中對應於該駕駛者之人臉的該眼部位置及/或該鼻子位置，獲得該駕駛者之人臉的一臉部轉向程度。 In step 623, the off-road tendency determination module obtains one of the driver's face based on the identified eye position and/or the nose position of the current image corresponding to the driver's face. The degree of face turning.
在步驟624中，該偏離車道傾向判定模組根據該當前影像所獲得之該駕駛者的該臉部轉向程度，標記該當前影像的一凝視狀態，該凝視狀態可為該普通狀態、一意圖右偏狀態及一意圖左偏狀態之其中一者。於本實施例中，當該臉部轉向程度指示出該駕駛者的臉部大致朝向前方，並未大幅度的轉向兩側的後照鏡或車內的後視鏡時，該凝視狀態為該普通狀態，值得一提的是，當該偏離車道傾向判定模組判定出無法自該當前影像辨識出對應於該駕駛者之人臉的該眼部位置及/或該鼻子位置時，該當前影像也會直接被標記為該普通狀態；當該臉部轉向程度指示出該駕駛者的臉部轉向右側的角度大於一第一預設角度時，該凝視狀態為該意圖右偏狀態，其中該第一預設角度可依實際情況調整；當該臉部轉向程度指示出該駕駛者的臉部轉向左側的角度大於一第二預設角度時，該凝視狀態為該意圖左偏狀態，其中該第二預設角度可依實際情況調整。此外，該第一預設角度不等於該第二預設角度，可根據車輛為左側駕駛座或右側駕駛座進行調整。於另一些實施例中，在根據該駕駛者的該臉部轉向程度標記該普通狀態、該意圖右偏狀態或該意圖左偏狀態時亦可同時考慮該駕駛者的臉部的轉向角度以及該眼睛的轉向角度，但不以此為限。 In step 624, the off-road tendency determination module marks a gaze state of the current image according to the degree of the face steering of the driver obtained by the current image, and the gaze state may be the normal state, an intended right One of a partial state and an intended left-biased state. In this embodiment, when the degree of steering of the face indicates that the driver's face is substantially facing forward, and the rear view mirror on both sides or the rear view mirror in the vehicle is not greatly turned, the gaze state is In the normal state, it is worth mentioning that when the off-road tendency determination module determines that the eye position and/or the nose position corresponding to the driver's face cannot be recognized from the current image, the current image It is also directly marked as the normal state; when the degree of steering of the face indicates that the angle of the driver's face turning to the right is greater than a first predetermined angle, the gaze state is the intended right-biased state, wherein the A preset angle may be adjusted according to an actual situation; when the degree of steering of the face indicates that the angle of the driver's face turning to the left is greater than a second predetermined angle, the gaze state is the intention left-biased state, wherein the The two preset angles can be adjusted according to actual conditions. In addition, the first preset angle is not equal to the second preset angle, and may be adjusted according to whether the vehicle is the left driver's seat or the right driver's seat. In other embodiments, the steering angle of the driver's face may also be considered simultaneously when the normal state, the intended right biased state, or the intended left deflected state is marked according to the driver's degree of face turning. The steering angle of the eye, but not limited to this.
在步驟625中，該偏離車道傾向判定模組根據該當前影像的該凝視狀態及先前所標記之該當前影像的前N張影像的N個凝 視狀態，分別計算該當前影像及該當前影像的前N張影像屬於該普通狀態、該意圖右偏狀態及該意圖左偏狀態的總數量。 In step 625, the off-road tendency determination module is based on the gaze state of the current image and the N marks of the first N images of the current image that were previously marked. Depending on the state, the current image and the first N images of the current image are respectively calculated to belong to the general state, the intended right deviation state, and the total number of the intended left bias states.
在步驟626中，該偏離車道傾向判定模組根據該普通狀態之總數量、該意圖右偏狀態之總數量及該意圖左偏狀態之總數量，獲得並傳送該駕駛者的該偏離車道傾向至該處理單元4。該偏離車道傾向指示出該主動朝右側偏離、該主動朝左側偏離或該未主動偏離車道等三種狀態。 In step 626, the deviation lane tendency determination module obtains and transmits the driver's deviation lane tendency according to the total number of the normal states, the total number of the intention right deviation states, and the total number of the intention left deviation states. The processing unit 4. The deviation from the lane indicates the three states of the initiative deviating to the right, the initiative deviating to the left, or the inactive deviating from the lane.
參閱圖1與圖4，在本實施例中，步驟626還包含以下子步驟。 Referring to Figures 1 and 4, in the present embodiment, step 626 also includes the following sub-steps.
在步驟627中，該偏離車道傾向判定模組判斷該意圖左偏狀態的總數量與該意圖右偏狀態的總數量中較大者是否大於或等於一預設值。當該偏離車道傾向判定模組判斷出該意圖左偏狀態的總數量與該意圖右偏狀態的總數量中較大者大於或等於該預設值時，流程進行步驟628。當該偏離車道傾向判定模組判斷出該意圖左偏狀態的總數量與該意圖右偏狀態的總數量中較大者小於該預設值時，流程進行步驟629。由於該人臉位置偵測單元1的影像拍攝模組有固定的取像頻率，若該偏離車道傾向判定模組計算該當前影像及該當前影像的前N張影像屬於該意圖左偏狀態的總數量與該意圖右偏狀態的影像總數量中較大者大於或等於該預設值，代表該駕駛者在一定時間區間內正頻繁地觀察左側或右側的車道狀況。 In step 627, the off-road tendency determination module determines whether the greater of the total number of intent left-biased states and the total number of intended right-biased states is greater than or equal to a predetermined value. When the off-road tendency determination module determines that the greater of the total number of the intended left-biased states and the total number of the intended right-biased states is greater than or equal to the preset value, the flow proceeds to step 628. When the off-road tendency determination module determines that the greater of the total number of the intended left-biased states and the total number of the intended right-biased states is less than the preset value, the flow proceeds to step 629. Since the image capturing module of the face position detecting unit 1 has a fixed image capturing frequency, if the off-track tendency determining module calculates the total number of the current image and the first N images of the current image belonging to the intended left-biased state The larger of the total number of images and the intended right-biased state is greater than or equal to the preset value, indicating that the driver is frequently viewing the left or right lane condition within a certain time interval.
在步驟628中，該偏離車道傾向判定模組以該意圖左偏狀態的總數量與該意圖右偏狀態的總數量中較大者所對應的偏離狀態作為該駕駛者的偏離車道傾向，並傳送該偏離車道傾向至該處理單元4。亦即，當該意圖右偏狀態的總數量大於該意圖左偏狀態的總數量時，該偏離車道傾向判定模組獲得指示出該主動朝右側偏離的該偏離車道傾向，當該意圖左偏狀態的總數量大於該意圖右偏狀態的總數量時，該偏離車道傾向判定模組獲得指示出該主動朝左側偏離的該偏離車道傾向。 In step 628, the off-road tendency determination module uses the deviation state corresponding to the larger of the total number of the intentionally left-biased states and the total number of the intended right-biased states as the deviation direction of the driver, and transmits This deviation from the lane tends to the processing unit 4. That is, when the total number of the intentionally right-biased states is greater than the total number of the intended left-biased states, the off-track tendency determination module obtains the off-track tendency indicating the active deviation toward the right side, when the intentional left-biased state The deviating lane tendency determination module obtains the off-track tendency indicating that the initiative deviates to the left side when the total number is greater than the total number of the intentionally right-biased states.
在步驟629中，該偏離車道傾向判定模組獲得指示出該未主動偏離車道的該偏離車道傾向，並傳送該偏離車道傾向至該處理單元4。 In step 629, the off-road tendency determination module obtains the off-track tendency indicating the non-active departure from the lane, and transmits the off-track tendency to the processing unit 4.
在本實施例中，該N值是以一預設秒數計算獲得，例如：於該步驟627中，欲以當前時間為基準判斷前3秒內屬於意圖左偏狀態或意圖右偏狀態是否超過該預設值，且該影像拍攝單元11每秒擷取60張影像，則該N值為3乘以60，即180。於此實施例中，該步驟627所選用的預設值可設定為例如：30。也就是說，當180張影像的標記結果中該意圖左偏狀態的總數量與該意圖右偏狀態的總數量中較大者大於30張時，則判定駕駛人欲主動地變換車道，接著，輸出該駕駛者的偏離車道傾向為該較大者。例如：該意圖左偏狀態的總數量大於該意圖右偏狀態的總數量且同時大於該預設值 時，輸出該駕駛者的偏離車道傾向為主動朝左側偏離至該處理單元4。 In this embodiment, the value of N is obtained by a predetermined number of seconds. For example, in step 627, it is determined whether the state of the intentional left-biased state or the intended right-biased state exceeds the first 3 seconds based on the current time. The preset value, and the image capturing unit 11 captures 60 images per second, the N value is 3 times 60, that is, 180. In this embodiment, the preset value selected in the step 627 can be set to, for example, 30. That is to say, when the total number of the intention left-biased states and the total number of the intended right-biased states are greater than 30 in the marking result of the 180 images, it is determined that the driver wants to actively change the lane, and then, The tendency of the driver to deviate from the lane is the larger one. For example, the total number of the intention left-biased states is greater than the total number of the intended right-biased states and is greater than the preset value at the same time. At the time, the driver's deviation from the lane is outputted to actively deviate to the left side of the processing unit 4.
在步驟63中，該處理單元4判斷是否接收到包含該車輛偏離方向的該車輛偏離資料。當該處理單元4判斷出接收到包含該車輛偏離方向的該車輛偏離資料時，流程進行步驟64。當該處理單元4判斷出未接收到包含該車輛偏離方向的該車輛偏離資料時，流程返回步驟61。在本實施例中，當該車輛偏離車道時，該車道偵測單元2產生包含該車輛偏離方向的該車輛偏離資料，並將該車輛偏離資料傳送給該處理單元4。 In step 63, the processing unit 4 determines whether the vehicle deviation data including the vehicle departure direction is received. When the processing unit 4 determines that the vehicle deviation data including the vehicle departure direction is received, the flow proceeds to step 64. When the processing unit 4 determines that the vehicle deviation data including the vehicle departure direction has not been received, the flow returns to step 61. In the present embodiment, when the vehicle deviates from the lane, the lane detecting unit 2 generates the vehicle deviation data including the vehicle deviation direction, and transmits the vehicle deviation data to the processing unit 4.
在步驟64中，該處理單元4判斷來自該人臉位置偵測單元1的該偏離車道傾向是否與該車輛偏離資料所包含的車輛偏離方向相同。值得一提的是，該處理單元4用以與該車輛偏離資料所包含的車輛偏離方向比較的該偏離車道傾向係最近一次自該人臉位置偵測單元1所接收到的該偏離車道傾向(亦即，該處理單元4用以與該車輛偏離資料所包含的車輛偏離方向比較的該偏離車道傾向係為對應有接收時間與判斷該偏離車道傾向是否相同於該車輛偏離方向的判斷時間間隔最短的偏離車道傾向)。當該處理單元4判斷出該偏離車道傾向與該車輛偏離方向相同時(亦即，該偏離車道傾向指示出該主動朝右側偏離且該車輛偏離方向也指示出該朝右側偏離，或該偏離車道傾向指示出該主動朝左側偏離且該車輛偏離 方向也指示出該朝左側偏離)，流程返回步驟61。當該處理單元4判斷出該偏離車道傾向與該車輛偏離方向不同時(亦即，該偏離車道傾向指示出該主動朝左側偏離或該未主動偏離車道，但該車輛偏離方向卻指示出該朝右側偏離，或該偏離車道傾向指示出該主動朝右側偏離或該未主動偏離車道，但該車輛偏離方向卻指示出該朝左側偏離)，流程進行步驟65。 In step 64, the processing unit 4 determines whether the off-track tendency from the face position detecting unit 1 is the same as the vehicle deviation direction included in the vehicle deviation data. It is worth mentioning that the deviation lane tendency of the processing unit 4 for comparing with the vehicle deviation direction included in the vehicle deviation data is the deviation lane tendency received from the face position detecting unit 1 last time ( That is, the deviation lane tendency of the processing unit 4 for comparing with the vehicle deviation direction included in the vehicle deviation data is the shortest determination time interval corresponding to the reception time and whether the deviation lane tendency is the same as the vehicle deviation direction. Deviation from the lane tendency). When the processing unit 4 determines that the deviation lane tendency is the same as the vehicle deviation direction (that is, the deviation lane tendency indicates that the initiative deviates to the right side and the vehicle deviation direction also indicates the deviation to the right side, or the deviation lane The tendency indicates that the initiative deviates to the left and the vehicle deviates The direction also indicates that the left side is offset), and the flow returns to step 61. When the processing unit 4 determines that the deviation lane tendency is different from the vehicle deviation direction (that is, the deviation lane tendency indicates that the initiative deviates to the left side or the vehicle does not actively deviate from the lane, but the vehicle deviation direction indicates the direction The deviation from the right side, or the deviation from the lane direction indicates that the initiative deviates to the right or the vehicle does not actively deviate from the lane, but the vehicle deviates from the direction indicating that the left side is deviated), and the flow proceeds to step 65.
在步驟65中，該處理單元4傳送一警示訊號至該警示單元3，以發出警示。 In step 65, the processing unit 4 transmits an alert signal to the alert unit 3 to issue an alert.
綜上所述，本發明車道偏離警示方法及其系統，藉由利用該駕駛者之影像來判定該偏離車道傾向與該車輛偏離資料所包含的車輛偏離方向是否同向可更加直覺且正確地判斷出該駕駛者是否為蓄意偏移車道。當該處理單元4接收到來自該車道偵測單元2的該車輛偏離資料且判定出該偏離車道傾向與該車輛偏離方向不同向時，才傳送一警示訊號至該警示單元3以發出該警示，藉此以避免該警示單元3誤報。此外，自該人臉位置偵測單元1獲得該駕駛者之偏離車道傾向，與自行車電腦取得行車資訊相較下，較為容易且方便，故確實能達成本發明的目的。 In summary, the lane departure warning method and system thereof of the present invention can determine whether the off-track tendency and the deviation direction of the vehicle included in the vehicle deviation data are more intuitive and correctly judged by using the image of the driver. Whether the driver is intentionally offsetting the lane. When the processing unit 4 receives the vehicle deviation data from the lane detecting unit 2 and determines that the deviation lane tendency is different from the vehicle deviation direction, an alert signal is transmitted to the warning unit 3 to issue the warning. Thereby, the warning unit 3 is prevented from being falsely reported. In addition, obtaining the driver's tendency to deviate from the lane from the face position detecting unit 1 is easier and more convenient than obtaining the driving information from the bicycle computer, so that the object of the present invention can be achieved.
惟以上所述者，僅為本發明的實施例而已，當不能以此限定本發明實施的範圍，凡是依本發明申請專利範圍及專利說明書內容所作的簡單的等效變化與修飾，皆仍屬本發明專利涵蓋的範圍 內。 However, the above is only the embodiment of the present invention, and the scope of the invention is not limited thereto, and all the simple equivalent changes and modifications according to the scope of the patent application and the patent specification of the present invention are still Scope covered by the patent of the present invention Inside.
Claims (6)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW106128035A TWI646001B (en) | 2017-08-18 | 2017-08-18 | Lane departure warning method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW106128035A TWI646001B (en) | 2017-08-18 | 2017-08-18 | Lane departure warning method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TWI646001B true TWI646001B (en) | 2019-01-01 |
| TW201912477A TW201912477A (en) | 2019-04-01 |
Family
ID=65803992
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW106128035A TWI646001B (en) | 2017-08-18 | 2017-08-18 | Lane departure warning method and system |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TWI646001B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010002953A (en) * | 2008-06-18 | 2010-01-07 | Mazda Motor Corp | Lane departure alarm device of vehicle |
| CN105151049A (en) * | 2015-08-27 | 2015-12-16 | 嘉兴艾特远信息技术有限公司 | Early warning system based on driver face features and lane departure detection |
| CN106467057A (en) * | 2015-08-21 | 2017-03-01 | 腾讯科技（深圳）有限公司 | The method of lane departure warning, apparatus and system |
-
2017
- 2017-08-18 TW TW106128035A patent/TWI646001B/en active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010002953A (en) * | 2008-06-18 | 2010-01-07 | Mazda Motor Corp | Lane departure alarm device of vehicle |
| CN106467057A (en) * | 2015-08-21 | 2017-03-01 | 腾讯科技（深圳）有限公司 | The method of lane departure warning, apparatus and system |
| CN105151049A (en) * | 2015-08-27 | 2015-12-16 | 嘉兴艾特远信息技术有限公司 | Early warning system based on driver face features and lane departure detection |
Also Published As
| Publication number | Publication date |
|---|---|
| TW201912477A (en) | 2019-04-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10445596B2 (en) | Camera device for vehicle | |
| US7532981B2 (en) | Vehicle departure detecting device | |
| US20190135291A1 (en) | System and method for controlling vehicle based on condition of driver | |
| US9542606B2 (en) | Lane line recognition apparatus | |
| KR101231510B1 (en) | System for alarming a danger coupled with driver-viewing direction, thereof method and vehicle for using the same | |
| CN106184232B (en) | A kind of lane departure warning control method based on driver visual angle | |
| CN113071415A (en) | Vehicle system and vehicle | |
| CN107111955B (en) | Operation control device | |
| JP2010049383A (en) | Warning device for vehicle | |
| JP2007072629A (en) | Onboard warning device | |
| JP4647387B2 (en) | Vehicle driving support device | |
| JP2009015533A (en) | Gaze direction detecting device | |
| US20180201276A1 (en) | Driver condition detection system | |
| JP2000127849A (en) | Rear and side monitor for vehicle | |
| JP2006331323A (en) | Vehicle deviation alarm device | |
| JP2010039919A (en) | Warning device | |
| KR101124948B1 (en) | Alarm system for informing start of front vehicle | |
| JP2008046845A (en) | Cutting-in vehicle determination apparatus | |
| JP2007249757A (en) | Warning device | |
| JP2007072570A (en) | Inattentive driving deciding device | |
| CN112119438A (en) | Vehicle rear-lateral-direction warning device and vehicle rear-lateral-direction warning method | |
| CN113895433B (en) | Forward collision avoidance system and method thereof through sensor angle adjustment | |
| JP2014229197A (en) | Vehicle notification device | |
| TWI646001B (en) | Lane departure warning method and system | |
| JP2008071062A (en) | Operation support device and notification method |