Method for predicting lane line and lane departure warning system using the same
First Claim
1. A method for reconstructing an unidentified boundary of a traffic lane including a left boundary and right boundary, comprising the steps of:
- capturing a plurality of traffic lane images by a front imaging unit;
providing a top boundary predicting relationship curve and a bottom boundary predicting relationship curve by a control unit, wherein the top boundary predicting relationship curve defines a top geometric relationship between the left boundary and the right boundary at top of a region of interest (ROI) in a traffic lane image, and the bottom boundary predicting relationship curve defines a bottom geometric relationship between the left boundary and the right boundary at bottom of the ROI in the traffic lane image, and the top boundary predicting relationship curve and the bottom boundary predicting relationship curve are determined by a determining process comprising the steps of;
identifying the left boundary and the right boundary of the traffic lane in the plurality of traffic lane images, and setting a virtual center line for the traffic lane;
forming the ROI in each traffic lane image, wherein each ROI in each traffic lane image comprises;
a top left point, wherein the top left point is where the left boundary intercepts top of the ROI;
a top right point, wherein the top right point is where the right boundary intercepts top of the ROI;
a bottom left point, wherein the bottom left point is where the left boundary intercepts bottom of the ROI; and
a bottom right point, wherein the bottom right point is where the right boundary intercepts bottom of the ROI;
acquiring a plurality of distance information sets by acquiring each distance information set from each of the traffic lane images, wherein each distance information set comprises;
a top left distance measured between the top left point and the virtual center line;
a top right distance measured between the top right point and the virtual center line;
a bottom left distance measured between the bottom left point and the virtual center line of the lane; and
a bottom right distance measured between the bottom right point and the virtual center line; and
performing a numerical analysis upon the plurality of distance information sets as a first set of characteristic values and a second set of characteristic values to obtain the top boundary predicting relationship curve and the bottom boundary predicting relationship curve;
reconstructing the unidentifiable boundary in any traffic lane image by applying the measured first and second set of characteristic values into the top boundary predicting relationship curve and the bottom boundary predicting relationship curve, by the control unit;
dynamically updating the top boundary predicting relationship curve and the bottom boundary predicting relationship curve when width of the lane is changed by repeating the determining process to update the top boundary predicting relationship curve and the bottom boundary predicting relationship curve;
detecting a vehicle status by a vehicle status detecting unit, and forwarding the vehicle status to control unit;
issuing an alarm signal by the control unit as the control unit evaluates the vehicle status and determine a precaution is needed; and
generating an alarm by an alerting unit, wherein the alerting unit is coupled to the control unit for receiving the alarm signal to generate the alarm.
1 Assignment
0 Petitions
Accused Products
Abstract
A method for predicting lane line is provided in the present invention, in which the method is capable of predicting and reconstructing the unidentified lane line according to a predetermined geometry relationship based on the recognized lane line data previously while one side of lane line is unable to be identified. In one embodiment, a lane predicting relationship formula is formed by describing the geometry relationship between the two identified lines defined the lane. When the one side of line of the lane is unable to be identified, a new predicted line can be reconstructed for representing the unidentified line of lane according to the lane predicting relationship curve. In another embodiment, a lane departure warning system utilizing the foregoing method is provided for monitoring vehicle moving status and alerting the driver so as to enhance the vehicle safety.
42 Citations
17 Claims
-
1. A method for reconstructing an unidentified boundary of a traffic lane including a left boundary and right boundary, comprising the steps of:
-
capturing a plurality of traffic lane images by a front imaging unit; providing a top boundary predicting relationship curve and a bottom boundary predicting relationship curve by a control unit, wherein the top boundary predicting relationship curve defines a top geometric relationship between the left boundary and the right boundary at top of a region of interest (ROI) in a traffic lane image, and the bottom boundary predicting relationship curve defines a bottom geometric relationship between the left boundary and the right boundary at bottom of the ROI in the traffic lane image, and the top boundary predicting relationship curve and the bottom boundary predicting relationship curve are determined by a determining process comprising the steps of; identifying the left boundary and the right boundary of the traffic lane in the plurality of traffic lane images, and setting a virtual center line for the traffic lane; forming the ROI in each traffic lane image, wherein each ROI in each traffic lane image comprises; a top left point, wherein the top left point is where the left boundary intercepts top of the ROI; a top right point, wherein the top right point is where the right boundary intercepts top of the ROI; a bottom left point, wherein the bottom left point is where the left boundary intercepts bottom of the ROI; and a bottom right point, wherein the bottom right point is where the right boundary intercepts bottom of the ROI; acquiring a plurality of distance information sets by acquiring each distance information set from each of the traffic lane images, wherein each distance information set comprises; a top left distance measured between the top left point and the virtual center line; a top right distance measured between the top right point and the virtual center line; a bottom left distance measured between the bottom left point and the virtual center line of the lane; and a bottom right distance measured between the bottom right point and the virtual center line; and performing a numerical analysis upon the plurality of distance information sets as a first set of characteristic values and a second set of characteristic values to obtain the top boundary predicting relationship curve and the bottom boundary predicting relationship curve; reconstructing the unidentifiable boundary in any traffic lane image by applying the measured first and second set of characteristic values into the top boundary predicting relationship curve and the bottom boundary predicting relationship curve, by the control unit; dynamically updating the top boundary predicting relationship curve and the bottom boundary predicting relationship curve when width of the lane is changed by repeating the determining process to update the top boundary predicting relationship curve and the bottom boundary predicting relationship curve; detecting a vehicle status by a vehicle status detecting unit, and forwarding the vehicle status to control unit; issuing an alarm signal by the control unit as the control unit evaluates the vehicle status and determine a precaution is needed; and generating an alarm by an alerting unit, wherein the alerting unit is coupled to the control unit for receiving the alarm signal to generate the alarm. - View Dependent Claims (2, 3, 4, 5)
-
-
6. A traffic lane departure warning system, comprising:
-
a front imaging unit, for capturing a plurality of traffic lane images of a traffic lane on which a vehicle travels on the traffic lane, wherein the traffic lane has a left boundary and a right boundary; a vehicle status detecting unit, for detecting a status of the vehicle and thus generating a status signal; an image processing/controlling module, reconstructing an identifiable boundary of the traffic lane and evaluating the status signal so as to generate an alarm signal; a displaying unit, coupled to the image processing/controlling module for displaying an image of lane line according; and an alerting unit, coupled to the image processing/controlling module for receiving the alarm signal and thus generating an alarm; wherein the image processing/controlling module reconstructs the identifiable boundary by; identifying the left boundary and the right boundary of the traffic lane in the plurality of traffic lane images, and setting a virtual center line for the traffic lane; forming the ROI in each traffic lane image, wherein each ROI in each traffic lane image comprises; a top left point, wherein the top left point is where the left boundary intercepts top of the ROI; a top right point, wherein the top right point is where the right boundary intercepts top of the ROI; a bottom left point, wherein the bottom left point is where the left boundary intercepts bottom of the ROI; and a bottom right point, wherein the bottom right point is where the right boundary intercepts bottom of the ROI; acquiring a plurality of distance information sets by acquiring each distance information set from each of the traffic lane images, wherein each distance information set comprises; a top left distance measured between the top left point and the virtual center line; a top right distance measured between the top right point and the virtual center line; a bottom left distance measured between the bottom left point and the virtual center line of the lane; and a bottom right distance measured between the bottom right point and the virtual center line; and performing a numerical analysis upon the plurality of distance information sets as a first set of characteristic values and a second set of characteristic values to obtain the top boundary predicting relationship curve and the bottom boundary predicting relationship curve. - View Dependent Claims (7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
Specification