JP2012181691A - Method for recognizing patten on road surface, and information recording device for vehicle - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for recognizing a pattern on a road surface and an information recording device for a vehicle, allowing recognition and preventing false recognition by using simple processing.SOLUTION: A method for recognizing a pattern on a road surface comprises the steps of: inputting Gray scale data of an image frame (S11); allocating a first processing target range in a first image frame (S13); scanning image data within the first processing target range in a predetermined direction to detect a combination of a plurality of patten edges each showing a specific linear pattern (S14); after once detecting the linear pattern, determining a second processing target range smaller than the first processing target range based on a position of the detected linear pattern (S16); scanning image data within the second processing target range in a second image frame in a predetermined direction to detect the linear pattern (S18); and when consecutively detecting the linear pattern a plurality of times, recognizing the linear pattern as an effective pattern (S21).

Description

  The present invention relates to a road surface pattern recognition method and a vehicle information recording device that can be used for a drive recorder or a digital tachograph mounted on a vehicle as an on-vehicle device, for example, and in particular, a white line or the like marked on a road surface of a road The present invention relates to a technique for recognizing a linear pattern.

  On general roads and highways, vehicles need to travel inside a predetermined travel lane (lane). Usually, the left and right end portions of each traveling lane are marked with a white line representing the boundaries of the plurality of traveling lanes, a center line representing the center of the road, a white line representing the end of the road, and the like. Therefore, the driver who drives the vehicle must check the position of the white line, center line, etc. with his / her eyes while driving carefully so that the driving position of the vehicle does not protrude from the current driving lane range. become.

  By the way, a recent vehicle may be equipped with an in-vehicle camera capable of photographing a landscape in front of or behind the vehicle. In-vehicle devices such as a drive recorder and a digital tachograph mounted on a vehicle can record video data obtained by photographing with the in-vehicle camera. A general drive recorder automatically records data such as images representing the situation before and after an abnormality such as a traffic accident on a recording medium. In addition, the digital tachograph records various data representing driving conditions periodically or as necessary.

  On the other hand, there is a possibility that a vehicle traveling on the road may continue to travel without departing from the traveling lane or perform meandering while wobbling from side to side due to, for example, carelessness of the vehicle driver or physical condition failure. Conceivable. Therefore, when a dangerous situation such as a deviation of a driving lane or a meandering operation occurs, it is desired that the vehicle-mounted device automatically alerts the driver. In addition, for example, in a company that owns business vehicles such as taxis and trucks, it is desirable that the administrator can grasp whether the driving situation of the crew of each vehicle is dangerous.

  By recognizing the white line representing the boundary of the driving lane based on the video taken by the above-mentioned on-board camera, the positional relationship between the white line and the host vehicle can be grasped. Detection is also possible.

  Conventional techniques disclosed in, for example, Patent Document 1 and Patent Document 2 are known as techniques for recognizing a white line by processing an image captured by an in-vehicle camera.

  In Patent Document 1, an edge extraction is performed on a captured image, a bird's-eye view is generated based on the edge-extracted information, and a white line recognition algorithm for recognizing a white line based on the bird's-eye view is disclosed in, for example, paragraphs “0011” and “0013”. Has been. Patent Document 2 mentions the possibility that the white line recognition algorithm is applied to a drive recorder.

Japanese Patent No. 3285575 Japanese Patent No. 4153522

  However, the conventional white line recognition algorithm requires very complicated and advanced processing for recognition. Therefore, in order to enable near real-time recognition, high-performance hardware is used for image data processing. It is unavoidable that the vehicle-mounted device becomes an expensive device. Further, for example, there is a possibility that noise such as an object existing on the road is erroneously recognized as a target pattern such as a white line, and in that case, the detection of the deviation from the driving lane or the meandering operation is adversely affected.

  The present invention has been made in view of the above-described circumstances, and an object of the present invention is to provide a pattern recognition method on a road surface that enables recognition by simpler processing than before and can suppress occurrence of erroneous recognition. Another object of the present invention is to provide a vehicle information recording apparatus.

In order to achieve the above-described object, the road surface pattern recognition method according to the present invention is characterized by the following (1) to (7).
(1) A road surface pattern recognition method for recognizing a linear pattern existing on a road surface based on image frame data obtained by continuously or intermittently photographing with a camera installed on a vehicle,
Input grayscale data of an image frame that includes gradation information representing the density of each pixel,
Assigning a first processing target range in the first image frame;
Scanning the image data of the first processing target range in a predetermined direction to detect a combination of a plurality of pattern edges representing a specific linear pattern;
After the linear pattern is detected once, a second processing target range having a size smaller than the first processing target range is determined with reference to the position of the detected linear pattern,
Among the second image frames taken at a different time from the first image frame, scan the image data in the second processing target range in a predetermined direction to detect the linear pattern,
Recognizing a corresponding linear pattern as an effective pattern when the linear pattern is continuously detected a plurality of times.
(2) The road surface pattern recognition method according to (1) above,
Before performing the detection of the linear pattern, the image data in the first processing target range and the second processing target range in the input image frame is coordinate-converted into contents suitable for image processing.
(3) The road surface pattern recognition method according to (2) above,
Generating information representing an image pattern arrangement on a plane that can be seen on a horizontal plane corresponding to the road surface in a state where the road surface is looked down from above by the coordinate conversion.
(4) The road surface pattern recognition method according to (3) above,
Before performing the coordinate conversion, edge information representing the contour of each pattern in the image is extracted from the input image, and coordinate conversion is performed on the extracted edge information.
(5) The road surface pattern recognition method according to (1) above,
As a combination of the plurality of pattern edges, a first edge representing a state in which the brightness of the image rapidly increases and a second edge representing a state in which the brightness of the image rapidly decreases appear within a predetermined distance. Use a combination.
(6) The road surface pattern recognition method according to (1) above,
When determining the position of the second processing target range, the movement distance of the pattern accompanying the movement of the vehicle between the time when the first image frame is captured and the time when the second image frame is captured is determined. The position calculated and shifted by the moving distance from the position where the linear pattern was previously detected is used as the reference position.
(7) The road surface pattern recognition method according to (1) above,
A color image signal is input from the camera, the signal is converted, and the grayscale data is generated for each image frame.

According to the road surface pattern recognition method having the configuration (1), the processing is relatively simple, the occurrence of erroneous recognition can be suppressed, and the processing can be performed without installing high-performance hardware. The time required for processing can be shortened by limiting the processing target range in the image frame. Moreover, a specific linear pattern can be recognized efficiently by detecting a combination of a plurality of pattern edges. Furthermore, by repeating the detection of the linear pattern a plurality of times for a plurality of image frames taken at different times, it is possible to prevent erroneous recognition of any noise that temporarily appears in the image as a linear pattern. . In addition, when detecting a linear pattern for the second time or later, the processing target range is determined based on the previously detected position, so that the area of the second processing target range can be reduced, and the processing burden can be reduced. .
According to the road surface pattern recognition method configured as described in (2) above, image transformation is performed as preprocessing before pattern recognition is performed, so that the burden of recognition processing can be reduced and the recognition processing is performed. The required time can be shortened.
According to the road surface pattern recognition method having the configuration of (3) above, the pattern arrangement image data in the horizontal plane equivalent to that on the road surface is processed when performing pattern recognition. Calculation becomes easy.
According to the road surface pattern recognition method having the configuration (4) described above, the information amount of the image to be coordinate-transformed can be greatly reduced, so that the processing burden can be greatly reduced.
According to the road surface pattern recognition method having the configuration (5), it is possible to easily detect a linear pattern such as a white line. That is, the white line representing the boundary of the driving lane appears in the image frame so as to extend in the vertical direction with a certain width, so when this image is scanned in the horizontal direction (the width direction of the road), one edge of both ends of the white line pattern The brightness increases rapidly, and the brightness decreases rapidly at the other edge. That is, the combination of the first edge and the second edge matches the feature of the white line pattern.
According to the road surface pattern recognition method configured as described in (6) above, it is possible to repeat the recognition of the same portion a plurality of times using images of a plurality of frames taken at different timings, and the influence of temporarily generated noise. Can be eliminated.
According to the road surface pattern recognition method configured as described in (7) above, image data suitable for edge detection processing of a linear pattern such as a white line is processed even when an image is captured using a color camera. Can be entered as

In order to achieve the above-described object, a vehicle information recording apparatus according to the present invention is characterized by the following (8).
(8) A function of inputting information including image frame data obtained by continuously or intermittently photographing with a camera installed on the vehicle and automatically recording the information on a predetermined recording medium. An information recording device for a vehicle,
Based on image frame data output from the camera, grayscale data of an image frame including gradation information representing the density of each pixel is input, and a first processing target range is assigned to the first image frame. The image data of the first processing target range is scanned in a predetermined direction to detect a combination of a plurality of pattern edges representing a specific linear pattern, and after the linear pattern is detected once, A second processing target range that is smaller than the first processing target range is determined with reference to the position of the detected linear pattern, and was captured at a time different from the first image frame. In the second image frame, when the linear pattern is detected by scanning the image data in the second processing target range in a predetermined direction, and the linear pattern is continuously detected a plurality of times, The Linear pattern that includes a road surface pattern recognition unit for recognizing a valid pattern.

  According to the vehicle information recording apparatus having the configuration (8), the road surface pattern recognition unit recognizes a linear pattern such as a white line on the road surface. It is also possible to automatically detect the situation, record this state and issue a warning.

  According to the road surface pattern recognition method and the vehicle information recording apparatus of the present invention, a necessary linear pattern can be recognized by a simpler process than before, and the occurrence of erroneous recognition can be suppressed.

  The present invention has been briefly described above. Further, details of the present invention will be further clarified by reading through the modes for carrying out the invention described below with reference to the accompanying drawings.

It is a flowchart showing the process sequence for recognizing the white line pattern on a road surface. It is a block diagram which shows the structural example of the hardware of the information recording device for vehicles. It is a schematic diagram showing the example of the recognition target image after coordinate transformation. It is a schematic diagram showing the relationship between a white line pattern and the detected edge pair.

  Specific embodiments relating to the road surface pattern recognition method and the vehicle information recording apparatus of the present invention will be described below with reference to the drawings.

  An outline of the processing procedure for recognizing the white line pattern on the road surface is shown in FIG. Moreover, the structural example of the hardware of the vehicle information recording device carrying the function which performs the process sequence of FIG. 1 is shown by FIG.

  The vehicle information recording apparatus 100 shown in FIG. 2 is assumed to be a drive recorder mounted on a vehicle as a vehicle-mounted device, but the present invention is applied to other vehicle-mounted devices such as a digital tachograph having the same function. Application is also possible.

  The vehicle information recording apparatus 100 shown in FIG. The in-vehicle camera 20 is fixed at a predetermined position on the vehicle, for example, near the center of the front windshield glass in the passenger compartment, near the driver's seat, near the rear window, etc. with the shooting direction facing the outside of the vehicle. The Accordingly, the in-vehicle camera 20 is similar to the scenery outside the vehicle seen from the viewpoint of the driver who drives the vehicle, the scenery in the forward or backward direction of the vehicle, that is, a part of the body of the own vehicle, the road surface, Other vehicles and pedestrians can be photographed continuously or intermittently. In this embodiment, the case where the camera which can image | photograph a color image as the vehicle-mounted camera 20 is assumed.

  Since the in-vehicle camera 20 is fixed on the vehicle, the landscape range captured by the in-vehicle camera 20 does not change when the vehicle does not move, but the landscape changes as the vehicle moves. For example, when a pattern such as a white line marked on the road surface is present, the positional relationship between the vehicle and the corresponding pattern changes as the vehicle moves, and the change appears as a change in the captured image. . Therefore, if a pattern such as a white line can be recognized in the captured image, a change in the position of the vehicle in the lateral direction within the traveling lane, a deviation from the traveling lane, a meandering operation, or the like It is also possible to detect such a situation.

  2 includes an image processing unit (road surface pattern recognition unit) 11, an acceleration (G) sensor 12, a main control unit 13, a non-volatile recording medium 14, a memory (SDRAM) 15, an interface (I / I). F) 16, 17 and 18 are provided.

  The image processing unit 11 is composed of elements mainly composed of a microcomputer (CPU), and realizes a predetermined function by executing a program prepared in advance. That is, the image processing unit 11 processes a video signal captured by the in-vehicle camera 20 to generate necessary image data, or performs processing for recognizing a predetermined pattern in the image. For example, the processing procedure illustrated in FIG. 1 is also executed by the image processing unit 11.

  The acceleration sensor 12 can detect the magnitude of acceleration applied to the vehicle. That is, the situation can be detected when a traffic accident occurs or when an abnormal state such as sudden acceleration or sudden deceleration occurs.

  The main control unit 13 is composed of elements mainly composed of a microcomputer (CPU), and realizes a predetermined function by executing a program prepared in advance. When an abnormal state such as a traffic accident occurs, the main control unit 13 performs a process of automatically recording image data and the like at timings before and after that. The main control unit 13 also has a function of detecting deviation from the traveling lane of the own vehicle based on the pattern recognized by the image processing unit 11, detecting meandering operation, and automatically recording these conditions. ing.

  The non-volatile recording medium 14 is a recording medium such as a memory card (for example, an SD card or a CF card) that is detachable from the drive recorder main body 10 or a hard disk device. That is, the nonvolatile recording medium 14 can freely write and read data, and can retain data contents even when the power supply is cut off.

  The memory 15 is a storage circuit in which data can be freely written and read, and is used to hold temporary data. For example, the memory 15 can be used as a buffer memory that holds image data for a certain period of time. Specifically, the image data obtained by the photographing by the in-vehicle camera 20 is always recorded on the memory 15 and overwritten with new image data over time-stale image data, from a predetermined time to the present. Past image data for a certain period of time can be held on the memory 15 with limited storage capacity.

  The vehicle speed sensor 21 detects the rotation of the output shaft of the transmission of the vehicle, and outputs one vehicle speed pulse as an electrical signal each time a predetermined amount of rotation is detected. By counting the number of detected vehicle speed pulses, it is possible to grasp the vehicle speed and the amount of movement of the vehicle.

  The image processing unit 11, the acceleration sensor 12, and the vehicle speed sensor 21 are connected to the main control unit 13 via interfaces 16, 17, and 18, respectively. The interfaces 16, 17, and 18 convert the outputs of the image processing unit 11, the acceleration sensor 12, and the vehicle speed sensor 21 to signals suitable for the processing of the main control unit 13, respectively.

Next, the contents of the processing procedure shown in FIG. 1 will be described.
In step S11, the image processing unit 11 takes in the latest image frame signal from the output of the in-vehicle camera 20 and converts it into a grayscale signal. That is, a gray scale signal suitable for image recognition processing, that is, a signal representing the density (gradation) of each pixel is generated from the color image signal output by the in-vehicle camera 20.

For example, by using the following equation, the luminance Y representing the gradation can be obtained for each pixel in the frame from the color signals of the three primary colors R (red), G (green), and B (blue).
Y = 0.2126 × R + 0.7152 × G + 0.0722 × B

  In step S12, edge extraction processing is performed using the grayscale data of the image frame input in step S11 as a processing target. That is, in order to extract a contour portion representing the feature of each pattern in the frame, each position on the frame is sequentially scanned using a predetermined filter, and the result is output. For example, edge extraction can be performed by processing an image using a known differential filter or sobel filter having a 3 × 3 pixel configuration corresponding to the pixel of interest and eight pixels adjacent thereto.

  In step S13, the image processing unit 11 determines a range for searching for a target pattern, and performs coordinate conversion of image data for this range. In this example, since the white line pattern representing the vehicle width direction limit position of the travel lane on the road is the recognition target, the search range is determined as follows in step S13.

Range in vehicle width direction (X direction): Range of reference position ± (Wr / 2) Range of travel direction (Y direction): Range of search length (Y1) determined in advance from the reference position Wr: Travel lane width (constant )
Reference position: for example, the camera mounting position in FIG.

  Further, the image data included in the search range is subjected to coordinate conversion to generate processing target image data as shown in FIG. 3, for example. In the video output from the in-vehicle camera 20, the influence of perspective appears, and the size and position of the pattern change depending on the distance between each pattern and the vehicle. Therefore, even for a linear pattern such as a white line, a relatively troublesome calculation must be performed in order to specify the actual position and range of the pattern.

  Therefore, as a pre-process before pattern recognition, coordinates are converted into an image with a simple calculation. That is, as shown in FIG. 3, the coordinates of each position of each pattern are converted so that it can be seen as a plan view in a state where each pattern is projected on a horizontal plane corresponding to the road surface of the road on which the host vehicle travels.

  FIG. 3 shows a state in which each pattern is projected on a plane composed of an X-axis corresponding to the vehicle width direction and a Y-axis representing the traveling direction of the vehicle, as in the state when the road is viewed from above. Therefore, for each of the white line patterns L1, L2, and L3 shown in FIG. 3, the influence of perspective is eliminated, and in the same manner as the shape of the actual white line pattern, a pattern with a constant width is arranged in parallel with each other. Appears.

  In step S14, the processing target image data generated in step S13 (data representing the pattern outline of the contents arranged as shown in FIG. 3) is scanned in order within the search range determined in step S13. Execute white line search algorithm.

  Specifically, the image data is scanned from the left to the right along the X-axis direction shown in FIG. 3, and when the scanning of one line is completed, the position in the Y-axis direction is shifted and again from the left to the right. Scan forward and repeat this process.

  By the way, when scanning the image data of the portion where the white line pattern exists in the X-axis direction, the brightness rapidly increases at the start edge position of the white line pattern, and the brightness decreases rapidly at the end edge position. Therefore, in the result of the edge extraction process in step S12, as shown in FIG. 4, for example, as shown in FIG. 4, positive edge information indicating that the brightness increases at the start edge position of the white line pattern (the differential value changes more than a predetermined value in the positive direction). ) Appears, and negative edge information (differential value changes more than a predetermined value in the negative direction) indicating a decrease in brightness appears at the end edge position of the same white line pattern.

  In general, since the width of the white line pattern displayed on the road surface is a predetermined constant value, the interval between the position where the positive edge information appears and the position where the negative edge information appears (in the X direction). Distance) is also almost constant.

  Therefore, in the white line search algorithm in step S14, when positive side edge information and negative side edge information appear at regular intervals as shown in FIG. 4, these combinations are used as edge pairs corresponding to the white line pattern. recognize.

  In step S15, it is identified whether or not an edge pair corresponding to the white line pattern has been detected as a result of the search in step S14. If not detected, the process returns to step S11. If detected, the process proceeds to the next step S16 in order to increase the recognition accuracy of the white line pattern.

  In step S16, a search range in the second and subsequent processes is determined. In this case, since the position where the candidate for the white line pattern exists has already been found in the first processing, it is limited to a range having a smaller area than the search range determined in step S13. Thereby, the amount of data to be processed can be reduced. In step S16, the search range is determined based on the position of the edge pair detected in step S14.

Specifically, the search range is determined as follows.
Range in the vehicle width direction (X direction): Range (Xa-Xm) to (Xb + Xm) Xa: Edge pair start position Xb detected last time: Edge pair end position Xm detected last time: Margin in X direction corresponding to error (margin) )
Range of traveling direction (Y direction): Range of (Ya + ΔY−Ym) to (Ya + ΔY + Ym) Ya: Y direction position of edge pair detected last time ΔY: Y direction movement amount corresponding to difference between previous and current images (vehicle speed and processing) (It depends on the shooting time difference of the target image frame)
Ym: margin in the Y direction corresponding to the error (margin)

  In step S17, the same processing as steps S11, S12, and S13 described above is performed except that the content determined in step S16 is adopted as the search range. That is, a newly captured image frame is captured from the in-vehicle camera 20, converted into grayscale data, subjected to edge extraction processing, and further coordinate conversion is performed to process a plan view as shown in FIG. Generate target data.

  In step S18, the processing target data generated in step S17 is searched within the search range determined in step S16 by the white line search algorithm similar to that in step S14 described above, and the aforementioned edge pair is detected.

  In step S19, it is identified whether or not an edge pair corresponding to the white line pattern has been detected as a result of the search in step S18. If not detected, the process returns to step S11. If detected, the process proceeds to the next step S20.

  In step S20, it is identified whether or not the same recognized white line pattern has been successfully recognized a predetermined number of times (in this example, three or more times including S14). If the number is less than the predetermined number, the process returns to step S16 to repeat the process, and if the number exceeds the predetermined number, the process proceeds to step S21.

In step S21, a pattern candidate corresponding to the edge pair detected in steps S14 and S18 is finally recognized as a white line pattern to be recognized.
That is, in the example shown in FIG. 1, when the same white line pattern detected at the same position on the road has been successfully recognized once in step S14 and further successfully recognized twice in step S18. As a white line pattern.

  Although not shown, after the white line pattern is determined in step S21, if the white line pattern at the corresponding position cannot be detected continuously in the three image frames, it is considered that the corresponding white line pattern has ended. The process is repeated from step S11.

  After the white line pattern is determined in step S21, the position of the white line pattern in the X direction is tracked in each image frame captured thereafter, thereby monitoring the position in the driving lane in the width direction and its change. It becomes possible to do. Therefore, for example, it is possible to detect the meandering operation of the host vehicle, to detect a departure from the traveling lane of the host vehicle, or to predict the occurrence of the departure before the departure occurs, and the main control unit 13 detects these This situation can be reflected in the recording operation of the drive recorder, or the driver can be warned in real time.

Note that the above-described embodiment is merely a representative example, and various modifications can be considered, for example, as described below when applied to an actual apparatus.
1. As for the camera that captures the scenery outside the vehicle, it is conceivable to add an optional camera in addition to the main camera (the above-described on-vehicle camera 20). In that case, a main camera and an optional camera are used simultaneously to shoot a plurality of images simultaneously. The main camera captures a relatively wide range, and the optional camera captures a relatively narrow range in the distance. In this way, the white line pattern at a location far away from the host vehicle can be recognized relatively easily.

  2. For example, by detecting the time zone at that time, detecting the brightness of the surroundings, analyzing the contents of the captured video, you can distinguish the difference between day and night, and identify whether you are traveling in the tunnel It becomes possible to do. In that case, if the difference between day and night and the difference in the status of whether or not it is in the tunnel are reflected in the recognition process of the white line pattern, the recognition accuracy of the white line pattern can be further improved.

  3. The in-vehicle camera 20, the image processing unit 11, and the interface 16 illustrated in FIG. 2 may be integrated. As a result, the structure of the apparatus can be simplified to reduce the cost, and the apparatus can be miniaturized.

  As described above, the road surface pattern recognition method and the vehicle information recording apparatus of the present invention can be assumed to be applied to, for example, a drive recorder or a digital tachograph mounted on a vehicle as an in-vehicle device. By applying the present invention, it is possible to provide a function for automatically recognizing a white line pattern without significantly increasing the cost of the apparatus, so that it is possible to detect a deviation from the driving lane of a vehicle or meandering operation that is a dangerous situation. Can be used to record information and output alarms. Further, not only a continuous solid white line pattern but also a dotted white line pattern can be recognized, and a solid line and a dotted line can be distinguished.

10 Drive recorder body 11 Image processing unit (CPU)
12 Acceleration sensor 13 Main control unit (CPU)
14 Non-volatile recording medium 15 Memory 16, 17, 18 Interface 20 In-vehicle camera 21 Vehicle speed sensor 100 Information recording device for vehicle Wr Travel lane width L1, L2, L3 White line pattern Y1 Search length

Claims (7)

A road surface pattern recognition method for recognizing a linear pattern existing on a road surface based on image frame data obtained by continuously or intermittently photographing with a camera installed on a vehicle,
Input grayscale data of an image frame that includes gradation information representing the density of each pixel,
Assigning a first processing target range in the first image frame;
Scanning the image data of the first processing target range in a predetermined direction to detect a combination of a plurality of pattern edges representing a specific linear pattern;
After the linear pattern is detected once, a second processing target range having a size smaller than the first processing target range is determined with reference to the position of the detected linear pattern,
Among the second image frames taken at a different time from the first image frame, scan the image data in the second processing target range in a predetermined direction to detect the linear pattern,
A road surface pattern recognition method characterized by recognizing a corresponding linear pattern as an effective pattern when the linear pattern is continuously detected a plurality of times. Before performing the detection of the linear pattern, the image data of the first processing target range and the second processing target range in the input image frame are coordinate-converted into contents suitable for image processing. The road surface pattern recognition method according to claim 1. The road surface pattern recognition method according to claim 2, wherein the coordinate conversion generates information representing an image pattern arrangement on a plane that can be seen on a horizontal plane corresponding to the road surface in a state where the road surface is looked down from above. . As a combination of the plurality of pattern edges, a first edge representing a state in which the brightness of the image rapidly increases and a second edge representing a state in which the brightness of the image rapidly decreases appear within a predetermined distance. A combination is used. The road surface pattern recognition method according to claim 1, wherein the combination is used. When determining the position of the second processing target range, the movement distance of the pattern accompanying the movement of the vehicle between the time when the first image frame is captured and the time when the second image frame is captured is determined. The road surface pattern recognition method according to claim 1, wherein a position obtained by calculating and shifting the linear pattern by the movement distance from a previously detected position is used as a reference position. The road surface pattern recognition method according to claim 1, wherein a color image signal is input from the camera, the signal is converted, and the grayscale data is generated for each image frame. Vehicle information having a function of inputting information including image frame data obtained by continuously or intermittently photographing with a camera installed on the vehicle and automatically recording the information on a predetermined recording medium A recording device,
Based on image frame data output from the camera, grayscale data of an image frame including gradation information representing the density of each pixel is input, and a first processing target range is assigned to the first image frame. The image data of the first processing target range is scanned in a predetermined direction to detect a combination of a plurality of pattern edges representing a specific linear pattern, and after the linear pattern is detected once, A second processing target range that is smaller than the first processing target range is determined with reference to the position of the detected linear pattern, and was captured at a time different from the first image frame. In the second image frame, when the linear pattern is detected by scanning the image data in the second processing target range in a predetermined direction, and the linear pattern is continuously detected a plurality of times, The The vehicle information recording apparatus characterized by comprising a road surface pattern recognition unit recognizes a linear pattern as a valid pattern.

Images