KR100490168B1 - The its Method and System for Lane departure warning using color data processing - Google Patents

Abstract

The present invention improves the recognition performance of lanes including color information as well as colored (yellow and blue) or black and white information which is important in a lane departure warning system through a camera of a lane departure warning device. It is to increase the efficiency.

In addition, it is an object of the present invention to quickly and in real time process the increased portion of the data amount due to the additional processing of the color information in the system configuration and processing in the above manner.

The main method is a lane boundary search algorithm for lane position recognition, which is a preprocessing stage of departure determination, which is a method of highlighting colored lanes, specifying a local area search for an effective lane, and maintaining lane positions within a specified search area. And a method for quickly detecting them. In addition, it provides a lane departure warning system through efficient memory management of the preprocessing process that processes the analog and digital image data, speed improvement of the postprocessing process, and optimization of the algorithm.

Description

Lane departure warning device using color information processing and the method {The its Method and System for Lane departure warning using color data processing}

In lane departure warning devices through image input means (eg cameras), black and white information is used to recognize lanes or markers, and the gray level between the general road and the lane on the road is divided based on the difference. It is a conventional technique.

In particular, in a black and white image, a yellow lane can hardly be distinguished from the road surface in a specific situation, or it is difficult to distinguish between a white lane and a colored lane. In the case of a yellow lane, the brightness difference is obvious even in a black and white image on the surface of a dark road such as an asphalt road. However, on bright road surfaces such as concrete surfaces, the brightness of the yellow lanes and the road surface are almost the same. In fact, when you observe black and white images of concrete roads, it is observed that yellow lanes do not appear.

In addition, when using this method, the distinction from the color of the colored lanes, such as yellow and blue, including white, is not accurately represented by the gray level in the actual road situation.

In this regard, the present invention is an important part in performance and function in a system such as a lane departure warning device that recognizes and warns whether a lane is detected through an image input means.

In light of the problems and situations described above, color information is additionally processed to recognize lanes on the road, thereby increasing the recognition rate of the yellow (center line) lanes, which are important in road driving conditions. We want to provide the driver with more accurate and stable alarms.

Hereinafter, with reference to the accompanying drawings will be described an embodiment of the present invention in more detail. 1 is a block diagram of a lane departure warning apparatus using color information according to an exemplary embodiment of the present invention.

The present invention provides a decoder unit (4) for receiving color image information data, separating and converting black and white information and color information, preprocessing processor units (6, 7) in charge of image signal processing thereof, and a lane departure warning device. Post-processing processor unit (1, 2, 3) that can further process the color information, the encoder unit for converting the color information to display color image information data using the lane-recognized result and the original image ( 5) consists of.

In addition, the preprocessor section is composed of a preprocessor (6) and preprocessor memory (7), the postprocessor is composed of a ROM (1) for the post-processor, a post-processor memory (2), a post-processor (3) .

The decoder unit is symmetrically converted into Y ₁ (8bit) + C _b (8bit) and Y ₂ (8bit) + C _r (8bit) in the size of 16bit with respect to the color image information data from the color camera, which is one pixel. Will be configured for Where Y ₁ and Y ₂ are the same value, C _b is the Blue value of the color information, C _r is the Red value of the color information, and other color values are Y ₁ or Y _2, or C _b. , CCIR601 format is determined by the combination of C _r .

Emphasizing the yellow component in the original image input means as shown in FIG. 2 for emphasizing the colored information in question in the black and white image; Differentially processing the original image with a differential filter to emphasize the vertical component; Designating a search area locally in the lane edge; Search for lane boundaries in the search area.

In the step of emphasizing the yellow component, when the yellow lane area is magnified in the concrete road image, the pixels are not exactly yellow. In other words, in the digital image, the yellow lanes gather various color dots to form a lane image. Therefore, the yellow lane highlighting effect cannot be obtained with the simple yellow highlighting alone. Therefore, the following processing is performed on the C _B color elements in the color space of Y, C _B , and C _R.

The following equation is applied to sum the color-enhanced image and the black-and-white image calculated by the above equation.

According to the above equation, a black and white image in which yellow lane areas are emphasized is obtained. The result obtained through the emphasis process on the color information is a result obtained by the emphasis process on the yellow lane, unlike the image in the black and white image.

Then, by performing a series of processes for recognizing a lane using only black and white image information, it is possible to obtain improved performance that can be processed more accurately and quickly.

As in the embodiment of the present invention, a method of emphasizing the yellow lane by converting a yellow value in a color image to a black and white level and adding it to the black and white level of the original pixel is specified from all color image information of other natural systems. By converting the color value of to black and white level, and adding it to the black and white level of the original pixel, the recognition rate of specific color information can be improved. Such a method can obtain improved performance than the result of only black and white image processing, and can obtain high-speed system performance similar to black and white image information processing time in processing more color image information than black and white image information.

The preprocessor 6 receives a function of writing image data of CCIR601 format into the preprocessor memory and a field synchronization (30 Hz) signal that is a division signal for one screen by receiving a synchronization signal (vertical synchronization, horizontal synchronization) output simultaneously from the decoder unit. It generates the function to perform alternate memory usage with the post-processor.

In addition, by using a filter (Fig. 3) to improve the stabilization and accuracy of the signal input from the external device, it performs a function to improve the performance of the system, and the synchronization signal (color sync, horizontal) required by the color encoder for video output. And a clock (LLC) for operation, and delivers the final output color image information in CCIR601 format.

The preprocessor memory 7 functions to store data of 256 x 192 pixels of size 16 bits.

The postprocessor ROM 1 instructs the postprocessor to start the operation of the postprocessor when the power is turned on using 8 bits or 16 bits as the algorithm for the post processor.

The post-processing processor 3 reads the color information data written in the pre-processor memory, and designates the local area where the lane of the previous frame is highlighted and the local area where the lane of the previous frame appears, that is, in the previous frame. The extracted lane equation is defined as the edge search interval of the current frame within 10 pixels of the left and right of the trajectory passing through the image plane. Vertical positions where the trajectory of the equation is outside the size range of the image are given a basic search range. The default search area allows you to find a new lane from the center of the screen when you cannot find a lane in the previous frame or when you change lanes.The signal comes from the vehicle's external ECU (turn left / right, brake, and speed signals from the flasher). It generates the output alarm when the input speed is over a certain speed and disables the output alarm when Turn left / right or brake signal occurs.

In addition, the lane recognition result obtained by emphasizing the color information, ECU signal information from the vehicle, unintended departure and accuracy, distance from the lane, distance from the lane of the own vehicle currently running, and road width through the preprocessor. Write to preprocessor memory.

The post-processor memory 2 is used to store temporary result values of color image information, vector tables of operation algorithms, and the like.

The encoder unit finally reads the result of the lane recognition extracted using the color information of the preprocessor and the postprocessor from the preprocessor 6, and the lane recognition result of the previous frame and the current frame of the current frame. OR operation of the color information data serves to output the final result to the display apparatus using the data and the synchronization signal to output the lane recognition result to the current image.

FIG. 5 is a result image after color information processing for the color sample image input shown in FIG.

FIG. 6 is a graph showing results of experiments on lane recognition rate using only existing black and white images under various conditions. FIG. 7 converts to black and white images after first performing color image information processing on a specific color according to the present invention. This is a result graph of the lane recognition rate obtained.

According to the present invention as described above, the yellow lane that has been a problem in the lane departure warning device that detects in advance the abnormal driving state of the vehicle or the unintentional lane departure state according to the driver's driving state and alerts in advance By improving the lane recognition rate for the public, a more stable system can prevent accidents and the corresponding situation in advance.

In addition, it is effective for the recognition of the object through the color information or similar aspects in the same form for industrial and industrial use.

On the other hand, the present invention is not limited only to the above-described embodiments, but may be modified and modified within the scope not departing from the gist of the present invention, such modifications and changes will be seen to belong to the following claims. .

1 is a block diagram of a lane departure warning device through yellow lane emphasis according to an embodiment of the present invention;

2 is a flowchart illustrating a lane departure warning method through yellow lane emphasis according to an embodiment of the present invention;

3 is a principle and configuration diagram of a filter for obtaining accurate timing and timing of an externally input signal in a preprocessor according to an embodiment of the present invention;

4 is an original image of color information data according to an embodiment of the present invention;

5 is a result image through color information emphasis processing according to an embodiment of the present invention,

6 is an experimental result graph showing a recognition rate of statistically recognizing lanes using only black and white information according to an embodiment of the present invention;

FIG. 7 is a graph showing results of a loss rate obtained by statistically recognizing lanes through colored lane emphasis through color information according to an exemplary embodiment of the present invention.

<Description of Symbols for Major Parts of Drawings>

1: ROM for post processor 2: Post processor memory

3: post-processing processor 4: color image decoder

5: Color Image Encoder 6: Preprocessing Processor

7: preprocessor memory

8: Internal system clock used internally by the preprocessor

9: aperiodic and sporadic input signal input from outside

10: signal delayed by one clock by the internal clock signal 9

11: Expression for generating NGT (Negative Going Transition) of external input signal and its result

12: Mathematical formula for generating positive moving transition of external input signal and its result

13: Characteristic result according to threshold applied to distinguish lane and road using only black and white information

14: Characteristic result according to the threshold applied to distinguish lane and road through colored lane emphasis through color information

Claims (8)

delete delete delete In the method for improving lane recognition using color image information, In the case of the yellow lane in the original color image, multiply this value by 3 by taking the absolute value of the difference between the values of the red component and the green to emphasize the yellow lane; Lane Recognition Method characterized by the process of summing the highlighted yellow lane image and black and white image In claim 4, Converting the yellow information value to black and white information value for color information highlighting, i.e. for C _B color elements in the color space of Y, C _B , C _R process; To sum the calculated color-enhanced and black-and-white images Lane recognition method characterized in that the conversion to the black and white image information by applying the emphasis process. delete delete In a system or apparatus that requires high-speed processing of color image information and improved recognition performance, A device using a method of variably improving the recognition rate for a specific color by converting the black and white level for a specific value to be distinguished and then emphasizing it in addition to the black and white level of the original pixel.