LANE DEPARTURE WARNING SYSTEM
Field of the Disclosure
The present disclosure describes an improved lane departure warning system extended to unstructured road boundary conditions.
Background of the Disclosure
In a conventional lane departure warning system, the lane recognition image processing apparatus are not capable of differentiating between different types of lane markings on the road such as continuous lane markings and broken lane markings. It also cannot identify the unstructured road boundaries such as kerb stones, crash barriers or particular sections of the road in which no edge of lane markings are provided and an unstructured boundary forms the edge of the lane. In such a case, the damage and the associated hazard, together with the resulting loss of confidence of the driver of the motor vehicle in the reliability of the lane departure warning system, will have an adverse effect on the road safety of the motor vehicle.
Another limitation of a conventional lane departure warning system is the type of image filter used to extract the images. When the forward view of the vehicle is taken by a camera installed thereon in a direction in which the vehicle is traveling, objects on an image thus taken become linearly smaller toward a vanishing point. Therefore, the width of the neighborhood of the pixel of interest to be referenced or viewed by an image filter is fixed, the actual width of an area extracted by the filter increases linearly in accordance with the increasing distance thereof from the camera. Accordingly, in case where a lane marking of a predetermined width on a road is detected as a physical quantity, the possibility of the presence of objects other than the lane marking becomes higher as the distance from the camera increases, resulting in reduced reliability in the result of the recognition of a distant portion of the lane marking.
Hence, a need exists to create an improved lane departure warning system that uses better image acquisition and image processing techniques.

Summary of the Disclosure
The present disclosure describes an improved lane departure warning system. It is in-vehicle; forward looking, vision-based electronic system for automobiles. It utilizes at least one camera installed on the front upper portion of the vehicle and takes the forward view of the road. The road regions are defined in terms of different cluster regions. As soon as a change is detected in the lane markings identified in the cluster region, an audio, a visual/video or a combination of both the signals is generated as a warning which shows that the vehicle is departing from its current lane in which it is present.
In another object of the present disclosure for the smooth functioning of the lane departure warning system an improved filtration mechanism is provided to extract the image which further assists the driver to maintain proper tracking.
Brief Description of the Drawings
The present disclosure explains the various embodiments of the instant invention with reference to the accompanying drawings present in the following description:
FIGURE 1 illustrates a block diagram showing the schematic construction of the Lane Departure Warning System according to an embodiment of the present disclosure.
FIGURE 2 illustrates a block diagram showing the image extraction and image filtering process according to an embodiment of the present disclosure.
Detailed Description of the Disclosure
The embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. However, the present disclosure is not limited to the embodiments and the embodiments described herein in the art of the present disclosure. The disclosure is described with reference to specific circuits, block diagrams, signals, etc. simply to provide a

more thorough understanding of the disclosure. The detailed description of the disclosure is as
follows:
Figure 1 illustrates a block diagram of a Lane Departure Warning System. A camera 101 is installed on a front upper portion of the vehicle and takes the forward view of the road. The real time image processing system 102 computes the clustered road region extraction and image filtering process. Depending upon the computation result three different audio-visual or audio or visual signals are generated by a signal generator 103 which are given by 104 a safe signal, 105 a lane change warning signal and 106 a lane change assist signal.
Further, the real time image processing means consists of a clustered road region extraction block 107, a filtration block 108, object detection block 109 and a data collection or storing block 110.
Figure 2 illustrates a block diagram showing the image extraction and image filtering process. As illustrated in figure 1 the lane departure warning system consists of at least one camera 101 which acquires the images from the forward view of the road. The camera is connected to the real time image processing means that computes the road region edge clusters in the step 201. Cluster forming involves combination of basic Edge Detection and Edge extension on the captured image. The edge detection is performed on the source image which has been detected using Canny Edge Detection Operator (developed by John F. Canny in 1986 and uses a multi-stage algorithm to detect a wide range of edges in images) with predefined low and high thresholds. Thus detected edges undergo a row-wise pixel-pair joining with predetermined and appropriate pixel distance threshold, which after complete frame processing results in the formation of binary regions.
The Binary regions image is then inverted which leaves the road, sky and other similar regions (with No texture and No change in the gray-level maintenance). A simple region filtration method based on its size (> [WIDTH/4 * HEIGHT/4]) and location will be performed to fix the final, featured road region.

The Lane detection is a two-phase procedure based on finding the uniqueness of a lane marking by combining all its features (size, shape, color information, geometry) which make it workable in all kinds of road scenarios. In its first phase, the algorithm uses the size, shape and color information of the lane and the geometry of the same in its second phase. Phase one combines the implementation of Binary-Blob formation and Horizontal gradient scan for the source image as parallel activities, and effectively uses the combined result for heavy noise reduction.
In general terms a Blob can be defined as a filled pixel area of same intensity or color. Binary Blob formation involves combination of basic Edge Detection and Edge extension on the edge detected image. Thus detected edges undergo a row-wise pixel-pair joining with a predetermined pixel distance threshold, which after complete frame processing results in the formation of binary regions called blobs.
The main idea behind this step of execution is the uniqueness for a lane marking in its intensity transition from the background to the fore ground. Hence the source image will undergo a Horizontal gradient scan in both X (Left to Right) and X' (Right to Left) directions, wherein at each pixel location, the intensity transition will be compared to a threshold with respect to its immediate left/right pixel location. The two-way scanned results will then be combined together to form all possible bright regions against a dark background.
In step 202 the common regions case of both Binary blob formation and Horizontal Gradient Scan are filtered out for its Phase-two processing. As per our calibration, consecutive two/three lane strips incline at similar angles with respect to a single point of view. The binary regions will then pass through an Angle-similarity-check for two iterations and thus filtered out regions can be marked as the featured lane strips. Coordinates of such lane strip regions will be updated and be used as a reference for the next frame.
In other embodiment of the disclosure when there is no boundary presents on the road then road region extraction can be executed in a step-by-step manner by Edge detection and Extension for Cluster formations, Image inversion for cluster ignorance, Size and location based

region filtration and a final road region approximation. Thus extracted road region image will undergo a vertical gradient scan (top to bottom) using Sobel operator and edge of the road region will be detected. In step 203 a Straight-line existence test (> predefined line length) will be performed on the edge map and a positive response updates of the boundary coordinates using the coordinates of the detected straight line. If no straight line is found in the edge map, then the edge-map will be approximated enough to update the road boundary coordinates. And this result is again processed for angle similarity check filtration. In the next iteration 204 all the lane point sequences are updated for every positive response.
Further, the computed results collected from the above detection and tracking three different signals are generated which are a safe signal 104 when there is no lane departure detected a lane warning signal 105 when there is a departure from the current lane and a lane assist signal 106 is generated if there is any object present in the nearby region of the lane.
The Lane Departure Warning system detects the lane in all environment condition if the lane is visible to camera and warns the driver of the vehicle if he departs from the current lane. It notifies the driver when lane markings are inadequate for detection or the system gets malfunctioned. It warns the driver of a lane departure when the vehicle is traveling above a certain speed and the vehicle's indicator signal is not activated. In addition to all above things, it is applicable in all personal/commercial vehicles (car, truck, passenger transport, heavy trucks etc.)
In the present disclosure, depending on the application, scope and criticality, various inputs from the car, including but not limited to indicator status and velocity input, may be taken in future developments.
Although the disclosure of the instant invention has been described in connection with the embodiment of the present disclosure illustrated in the accompanying drawings, it is not limited thereto. It will be apparent to those skilled in the art that various substitutions, modifications and changes may be made thereto without departing from the scope and spirit of the disclosure.

We Claim:
1. A Lane Departure Warning System, the system comprising
at least one image acquisition means (lOl)for capturing the image of the road;
a real time image processing means (102) including an extraction means for extracting the desired road image area from the image based on the predetermined values; and the processor for processing the extracted image with respect to the vehicle utilizing object detection and lane boundaries extraction mechanism; and
a tracking means (103) for detecting the state of change and sending a warning signal to an output means.
2. The Lane Departure Warning System as claimed in claim 1, wherein the image acquisition means (101) is a camera installed on the front upper portion of the vehicle and takes the forward view of the road.
3. The Lane Departure Warning System as claimed in claim 1, wherein the real time
processing means (102) comprises
a first extraction means (107) for extraction of cluster based road regions; and a filtration means (108) for producing filtered images.
4. The Lane Departure Warning System as claimed in claim 3, wherein the first extraction means uses edge extension and detection by the techniques of binary-blob formation and horizontal gradient scan.
5. The Lane Departure Warning System as claimed in claim 3, wherein the filtration
means uses gradient filters to compute the filtered value by combining the result of binary-blob formation and horizontal gradient scan.

6. The Lane Departure Warning System as claimed in claim 1, tracking means includes a
signal generator mechanism for generating an output signal,
7. The Lane Departure Warning System as claimed in claim 6 wherein said signal
generator mechanism generates any one of the following output signals
a safe signal (104) is generated when there is no lane departure detected;
a lane warning signal (105) is generated when there is a departure from the
current lane; and
a lane assist signal (106) is generated if there is any object present in the nearby
region of the lane.
8. The Lane Departure Warning System as claimed in claim 7, wherein the signals (104, 105, 106) can be an acoustic signal or a video signal or a combination of both.
9. A method for Lane Departure Warning, the method comprising
capturing the image of the road;
extracting the desired road image area from the image based on the predetermined values; and processing the extracted image with respect to the vehicle utilizing object detection and lane boundaries extraction mechanism; and
tracking and detecting the state of change and sending a warning signal to an output means.