JP2007128141A - System and method for determining road lane number in road image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system and method for determining the number of lanes by identifying each lane of a road part in a ground surface image obtained by photographing the ground surface from an artificial satellite or airplane. <P>SOLUTION: The road lane number determination system for determining the number of lanes of a road contained in the ground surface image obtained by photographing the ground surface from the artificial satellite or air plane comprises a lane boundary detection means having at least ground surface image data and road vector data, and detecting a lane boundary contained in ground surface image data corresponding to each road vector, and a road lane number determination means determining, based on information of the detected lane boundary information, the number of lanes for each road vector. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a system and method for identifying each lane of a road portion and determining the number of lanes in a ground image obtained by photographing the ground surface from an artificial satellite or an aircraft.

  In recent years, the use of artificial satellite images or aerial images (hereinafter referred to as “surface images”) obtained by photographing a ground surface from a high place with artificial satellites or aircraft (hereinafter referred to as “artificial satellites”) has become widespread. Yes. For example, it is used for the purpose of creating a map, grasping the use situation of land in an urban area, or analyzing the existence state and growth state of trees in a forest or the like. With the improvement of the resolution of the imaging device, it has become possible to acquire ground image data of a vast area on the ground at high speed.

  On the other hand, the use of a car navigation system using a GPS (Global Positioning System) that grasps the current position by receiving radio waves from an artificial satellite (GPS satellite) and measuring the latitude and longitude is becoming more and more widespread. Some recent car navigation systems display not only road maps but also road signs and lane information, which assists the driver in driving.

  Conventionally, road map data used for checking a route to reach a certain destination is configured as a symbol map including road vector information indicating which point is connected to which point by a road. However, with the recent development of car navigation systems, it is not limited to the display of route information from the starting point to the destination, and it is desired to realize various functions for improving the convenience of the driver. There is a lack of information on the map. In response to such changes, many road maps have been updated from conventional vector maps to detailed road maps. In order to increase the usefulness of information, the road map needs to be updated frequently.

  Conventionally, in order to create such a detailed road map, the position of the road is read from an existing vector road map, and the actual vehicle is loaded with cameras and sensors, and the road is traveled. The road data is created based on the information. For example, information such as road lanes and lane widths can be grasped in detail by a vehicle loaded with cameras and sensors as described above traveling in each lane on the road. Patent Document 1 describes a road map creation system using such a technique.

JP 2002-318533 A. Masahide Hirabayashi, “Latest Program Dictionary in C Language, Volume 2”, Technical Reviewer, November 27, 1992, p. 104-110 Masami Kaneko and five others, “Study on remote sensing method for marsh vegetation classification” Zhao, T. and Nevatia, R. “Car detection in low resolution aerial image”

  However, in order to create detailed road data using the above-described conventional technology, it is necessary to actually travel on all roads existing on the map in order to create / update the detailed road map. This entails enormous costs for running and measuring, and scheduling to create an efficient route to cover all roads (and this is based on incomplete road information before the update). (Which must be done) is very cumbersome and adds extra time and cost to roads that are congested or tolled.

  The present invention has been made in view of such circumstances, and intends to provide a system and method capable of automatically identifying each lane of a road portion and determining the number of lanes from ground image data. is there.

  As a result of earnest research in view of the above problem, the present inventor obtained from an artificial satellite or the like by utilizing the fact that the lane boundary lines that divide the lane on the road are regularly arranged in a regular shape. We have come up with the idea that lane boundaries can be automatically identified from the ground image and the number of road lanes can be estimated.

  That is, the present invention is a system for determining the number of road lanes included in a ground image obtained by photographing a ground surface from an artificial satellite or an aircraft, and has at least ground image data and road vector data. Lane boundary detection means for detecting a lane boundary included in the corresponding ground image data, and lane number determination means for determining the number of lanes for each road vector based on the information of the detected lane boundary. The present invention provides a road lane number determination system characterized by comprising the above.

  The present invention is also a system for determining the number of road lanes included in a ground image obtained by photographing a ground surface from an artificial satellite or an aircraft, and has at least ground image data and road vector data. Exclusion region determination means for determining an exclusion region that does not detect a lane boundary, and lane boundary detection that detects a lane boundary included in a region other than the exclusion region of the ground image data corresponding to each road vector And a road lane number determination system, characterized in that the road lane number determination system includes a lane number determination means for determining the number of lanes based on information of the detected lane boundary line for each road vector. is there.

  The present invention is also a system for determining the number of road lanes included in a ground image obtained by photographing a ground surface from an artificial satellite or an aircraft, and has at least ground image data and road vector data. Exclusion region determination means for determining an exclusion region that does not detect a lane boundary line, and a lane boundary line included in a region other than the exclusion region in the ground image data corresponding to a road vector having a predetermined length or more. Lane boundary detecting means for detecting, lane number determining means for determining the number of lanes based on information of the detected lane boundary line for each road vector, and a road vector not having the predetermined length or more A vehicle that estimates the number of lanes for the road vector based on the number of lanes determined for one or more other road vectors in the vicinity of the road vector There is provided a road lane number determining system characterized by comprising a number estimating means.

  In the system for determining the number of road lanes according to the present invention, the system for determining the number of road lanes according to the present invention further comprises color format conversion means for converting ground image data into HLS format and RGB format, and the exclusion area determination The means and the lane boundary line detecting means perform exclusion area determination and lane boundary line detection using one or both of ground image data in HLS format and RGB format.

  In the system for determining the number of road lanes according to the present invention, the lane boundary detection means performs lane boundary detection using data obtained by binarizing the brightness of HLS format ground image data. Usually, the lane boundary line is conspicuously colored in a completely different color from the road surface, so that it can be easily distinguished by comparing the brightness of both.

  In the system for determining the number of road lanes according to the present invention, the lane boundary detection unit is configured to determine the lane boundary based on the information of the lane boundary detected around the exclusion area for the exclusion area in the ground image data. It is characterized by estimating the presence of a boundary line. Normally, lane boundaries are regularly arranged at regular intervals in the vertical and horizontal directions of the road, so even if there is an area where the lane boundary could not be detected in part of the ground image Considering the above regularity, it is possible to interpolate the lane boundary line that should exist in the exclusion region.

  In the system for determining the number of road lanes according to the present invention, the lane boundary detection means divides ground image data corresponding to each road vector by a predetermined width along the road vector length direction, and each divided image data is A lane boundary line is detected by determining whether or not a lane boundary line is included. Usually, the lane boundary line has a long and narrow rectangular shape in the traveling direction of the road, so that the image division facilitates detection. It is preferable that the division width is approximately the same as a general lane boundary line.

  In the system for determining the number of road lanes according to the present invention, the lane boundary detection means is an image area having a pixel value or brightness different from that of other areas in the divided image data, and is a predetermined value or more in the road vector length direction. A region having a length of 2 is detected as a lane boundary line. This is because the lane boundary line usually has a certain length in the road traveling direction, and it is difficult to consider that an image area having a length less than this is a lane boundary line.

  In the system for determining the number of road lanes according to the present invention, the exclusion area determination means determines at least one of an intersection area, a vehicle, a vegetation, and a median strip included in the ground image data as an exclusion area. To do. In addition to these, as long as it can be distinguished from the road area based on the appearance characteristics of the object, this can be detected and set as an excluded area.

  In the system for determining the number of road lanes according to the present invention, the lane number estimation means, for road vectors not having a predetermined length or more, of the other road vectors connected from the respective ends of the road vectors, Two road vectors that have been determined and are closest to the respective ends are detected, and the number of lanes for the road vector is determined by setting the number of lanes determined for each of the two detected road vectors as the upper and lower limits. It is estimated that it is a numerical value within the range. Normally, except for the vicinity of the intersection, the number of lanes on the road does not change significantly, and thus such estimation can be performed.

As described above, according to the present invention, surface image data obtained from an artificial satellite or the like without having to travel and measure all the roads existing on the map with a lot of labor and cost. There is provided a system and method that can automatically identify each lane of a road portion and determine the number of lanes using.
The number of road lanes on the ground image corresponding to each road vector can be recognized.

  Hereinafter, the best mode for carrying out the system and method for determining the number of road lanes of the present invention will be described in detail with reference to the accompanying drawings. 1 to 15 are diagrams illustrating embodiments of the present invention. In these drawings, the same reference numerals denote the same components, and the basic configuration and operation are the same. To do.

System Configuration FIG. 1 is a functional block diagram schematically showing the configuration of a road lane number determination system according to this embodiment. In FIG. 1, this system is a storage device comprising a processing device 10 composed of a personal computer, a workstation, etc., a RAM (Random Access Memory) used as a main memory, and an auxiliary storage device such as a magnetic disk storage device. 20 and an input / output device 30.

  The input / output device 30 includes an input device 31 including a pointing device such as a keyboard and a mouse, a display device 32 such as a CRT display device, and a printer 33. The input device 31 is used for inputting parameters by the user, starting commands, and the like. The display device 32 and the printer 33 are used for presenting the determination result of the number of road lanes by this system to the user. The display device may include only one of the display device 32 and the printer 33.

  The storage device 20 stores ground image data 21, road vector data 22, HLS data 23, vegetation data 24, search exclusion area data 25, partial road lane data 26, and road lane data 27. Of these, the ground image data 21 and the road vector data 22 are data stored in advance before the processing by the present system is executed. The surface image data 21 is obtained as an artificial satellite image or an aerial image, and is RGB, IR (near infrared) 4-band multi-value image data. The road vector data 22 includes data indicating the positions of the start and end points of the road and the width of the road, and can be used as conventional road map data. A specific data configuration example of the road vector data 22 is as shown in FIG.

  The HLS data 23, the vegetation data 24, the search exclusion area data 25, the partial road lane data 26, and the road lane data 27 stored in the storage device 20 are subjected to predetermined processing on the ground surface image data 21 and the road vector data 22 ( It is generated as a result of performing (described later).

  The road lane number determination program 40 includes a color format conversion unit 100, a search exclusion area detection unit 200, a partial road lane detection unit 300, and an undecided road lane estimation unit 400. These are composed of a program module or the like that describes a procedure for performing processing on various data stored in the storage device 20, and details of each processing will be described later.

Details of Processing for Determining the Number of Road Lanes Next, details of the processing for determining the number of road lanes included in the ground image in this system will be described. Currently, the resolution of available ground images is on the order of tens of centimeters, even with the highest accuracy, while lane boundaries on roads are usually as large as several tens of centimeters in width and several meters in length. That's it. Therefore, in many cases, the resolution of the ground image is a critical resolution for recognizing the lane boundary. Therefore, this system is characterized in that appropriate detection processing is performed in consideration of the shape, arrangement, etc. peculiar to the lane boundary on the road so that the fine lane boundary can be easily detected from the ground image. .

  In this system, when the road lane number determination program 40 is started, the color format conversion unit 100, the search exclusion area detection unit 200, the partial road lane detection unit 300, and the undecided road lane estimation unit 400 are executed in this order. . The details of the processing by each will be described below.

(1) Color Format Conversion Processing The color format conversion unit 100 reads the ground image data 21 and performs color format conversion conversion. The surface image data 21 captured by an artificial satellite or the like usually has information of four bands of RGB and IR, but the color format conversion unit 100 converts the RGB color image data into a hue (H: Hue), The image data is converted into HLS space image data composed of lightness (L: Lightness) and saturation (S: Saturation). FIG. 3A shows an RGB color model. As a model for associating the RGB space with the HLS space, a bihexagonal pyramid color model shown in FIG. 3B is used. Here, H is a hue angle, and is specified by a value of 0 to 360 degrees. Red is 0 degrees, yellow is 60 degrees, green is 120 degrees, light blue is 180 degrees, blue is 240 degrees, purple is 300 degrees, and it returns to red again at 360 degrees. L is lightness, and is specified by a real value between 0 and 1. The 0 side is dark and the 1 side is bright. 0.5 is intermediate. The 0 to 0.5 side is a state where black paint is added, the 0.5 to 1 side is a state where white paint is added, and 0.5 is a state where neither is added. The lightness value represents the brightness component of the color image and matches the monochrome image information generated from the color image information. For detailed algorithms of RGB → HLS conversion and HLS → RGB conversion, see Non-Patent Document 1 and the like. The HLS image data converted by the color format conversion unit 100 is stored in the storage device 20 as HLS data 23.

  When the ground image is handled in the subsequent processing, either the ground surface image data 21 or the HLS data 23 is appropriately used. Which format data is used in each process may be determined from the viewpoint of detection accuracy and calculation amount.

  Further, the color format conversion unit 100 calculates the vegetation index from the R and IR values of the ground image data 21, and creates the vegetation data 24. The vegetation data 24 is data indicating a portion where the road of the ground image is covered with trees or the like. Here, the vegetation index is an index designed to represent the amount and activity of vegetation from the spectral reflection characteristics of vegetation. Various vegetation indices have been proposed so far, but this system uses the most commonly used NDVI (Normalized Difference Vegetation Index) (see Non-Patent Document 2). NDVI is defined by the following equation.

このようにして計算された植生指数ＮＤＶＩが植生データ24となる。

The vegetation index NDVI calculated in this way becomes the vegetation data 24.

(2) Search Exclusion Area Detection Processing Next, the search exclusion area detection unit 200 calculates an area that becomes noise in the road portion in the ground image, generates search exclusion area data 25 indicating the area, and stores it in the storage device 20. Store. FIG. 4 is a flowchart showing the flow of processing by the search exclusion area detection unit 200. As illustrated in FIG. 4, the search exclusion area detection unit 200 sequentially executes each subroutine of an intersection area detection process 201, a vehicle noise detection process 202, a vegetation noise detection process 203, and a search exclusion area determination process 204. Details of these processes will be described below.

  In the intersection area detection process 201, an area that is an intersection of roads in the ground image is detected. This intersection area is to be a search exclusion area in the search exclusion area determination processing 204 described later. Since the lanes intersect at the intersection, the road pattern is different from that of normal roads. In order to detect the lanes of the partial roads by the partial road lane detector 300 described later, such image data is excluded. It is necessary. For example, as shown in FIG. 5, image data obtained by binarizing an intersection area and another road portion in the ground image may be generated and temporarily stored.

  In the vehicle noise detection process 202, an area occupied by the vehicle in the road portion of the ground image is detected (refer to Non-Patent Document 3 for this detection method). The area occupied by this vehicle becomes noise when detecting the pattern of the lane boundary line on the road, and therefore should be set as the search exclusion area in the search exclusion area determination processing 204 described later. For example, as shown in FIG. 6D, image data obtained by binarizing an area occupied by vehicles in a road portion in the ground image and other road portions may be generated and temporarily stored.

  In the vegetation noise detection processing 203, a vegetation part in a road part in the ground image is detected. This processing can be performed using the vegetation data 24 already generated by the color format conversion unit 100. Vegetation existing on the road includes, for example, mediation of median strips and roadside trees that cover the road from above. In the area determination process 204, it should be a search exclusion area. For example, as shown in FIG. 6C, image data obtained by binarizing a vegetation portion and another road portion in a road portion in the ground image may be generated and temporarily stored.

  In the search exclusion area determination process 204, the intersection area, vehicle noise, and vegetation noise that should be excluded from the ground image are stored in the storage device 20 as the search exclusion area data 25 based on the results of the above-described processes 201 to 203. . For example, a logical sum of the binary image data generated by the above processes can be taken as a search exclusion area. FIG. 6 is a diagram illustrating an example in which various noises (c) to (e) are detected from the image (a) of the road portion in the ground image. In addition to the above, for example, when an area with a certain width that is different in color or brightness from other road parts is detected in the vicinity of the center of the road, this is determined as a median strip, You can also

(3) Partial Road Lane Detection Processing Next, the partial road lane detection unit 300 performs processing for detecting the lane boundary line for each road vector and estimating the number of lanes for the road portion included in the ground image. FIG. 7 is a flowchart showing the flow of processing by the partial road lane detection unit 300. In FIG. 7, the partial road lane detecting unit 300 first picks up one unprocessed road vector (step 301), and compares the length of the road vector with the preset minimum length of the road vector. (Step 302). If the length of the road vector is shorter than the minimum length, it is determined that information necessary for processing is insufficient, and the road vector is temporarily stored as an undetermined road vector (step 310). With respect to the undetermined road vector, the number of lanes can be estimated by an undetermined road lane estimation unit 400 described later.

  For a road vector having a sufficient length, the road portion of the ground image corresponding to the road vector is divided into columns having a width of a preset lane boundary parallel to the road vector length direction (step 303). ). The divided columns are selected one by one (step 304), and the brightness is binarized (step 305). However, the search exclusion region included in the column is excluded from the binarization target as noise. It is determined whether or not a lane boundary line is included in the binarized column image (step 306). The processes in steps 304 to 306 are performed on each divided column (step 307).

  In general, since the road surface and the lane boundary line have different brightness, the road surface and the lane boundary line can be distinguished by binarizing the road image. Various well-known techniques can be applied to the binarization of the image, but this system uses a discriminant analysis method. In the discriminant analysis method, when the histogram of the brightness of the entire target image is divided into two classes (classes) at a certain value ft, the variance between classes / (class variance within class 1 + class variance within class 2) ) Is determined as the threshold value when ft is the maximum.

  When binarization is performed on an image, it is often performed on the entire image. However, if the distribution of the brightness of an object is different for each part of the image, binarization cannot be performed appropriately. For example, since the distribution of lightness is different between an asphalt road and a concrete road, the appropriate lightness threshold value is different, and if the lane boundary line is also different in color, the appropriate lightness threshold value is different. Therefore, in this processing, in order not to determine images with mixed road surfaces and lane boundary lines of different colors, the road image is divided into columns with a width of about the lane boundary line parallel to the lines, and On the other hand, binarization is performed and lane boundary determination is performed. As shown in FIG. 8, the brightness histogram is clearly divided into two classes by this method, as shown in FIG. 8, so the lane boundary line and the road surface are separated by binarizing the image. It can be easily distinguished.

  FIG. 9 is a flowchart showing details of the lane boundary determination process in step 306. As a result of the processing so far, the binarized images of the divided columns can be classified into three types of regions, that is, a bright region, a dark region, and a search exclusion region, as shown in FIG. 6B. . In FIG. 9, the partial road lane detection unit 300 selects bright areas one by one from the row image and sequentially checks them (Step 501). It is determined whether or not the selected bright area is adjacent to the search exclusion area (step 502). If the selected bright area is adjacent to the search exclusion area, it is determined as an undetermined bright area where it cannot be determined whether or not the area is a lane boundary line ( Step 506). If the selected bright area is not adjacent to the search exclusion area, the length of the bright area is checked (step 503). It is determined whether or not the length of the bright region is equal to or greater than the minimum length for the preset lane boundary line (step 503). If the length of the bright area is equal to or greater than the above minimum length, the bright area is temporarily stored as a lane boundary line (step 504), and the length of the bright area is less than the above minimum length. Temporarily stores the bright area as a non-lane boundary line (step 505). The processing in the above steps 501 to 506 is performed for all the bright regions included in the row image (step 507). Thus, in the lane boundary determination process in step 306, the bright areas included in the column image are classified into one of three types: lane boundary lines, non-lane boundary lines, and undecided bright areas.

  In FIG. 7 again, after performing the above-described lane boundary determination processing for all the columns of the corresponding image of the selected road vector, the partial road lane detection unit 300 performs the above-described undecided bright region. Processing for estimating is performed (step 308). Here, as an example, a case where an undetermined bright region is estimated in the determination result shown in FIG. 11 obtained by the lane boundary determination processing in step 306 will be described. In FIG. 11, the length direction of the road vector is the Y direction, and the direction perpendicular thereto is the X direction.

  FIG. 10 is a flowchart showing details of the undetermined bright region estimation process in step 308. In FIG. 10, the partial road lane detecting unit 300 selects one undecided bright area in the corresponding image of the selected road vector (step 601), and in the road in the same traveling direction as this undecided bright area (road There are more than a certain number of lane boundaries in the vector-corresponding image (the area on the same side of the center line) in the area where the undecided bright area is almost the same in the X direction and the area where the Y direction is almost the same It is determined whether or not to perform (step 602). If there are more than a predetermined number of lane boundary lines, this undecided bright area is temporarily stored as a lane boundary line (step 603). If a predetermined number or more of lane boundary lines do not exist, this undecided bright area is temporarily stored as a non-lane boundary line (step 604). The processes in steps 601 to 604 described above are performed on all the undecided bright areas included in the corresponding image of the selected road vector (step 605). In this undecided bright region estimation process, as shown in FIG. 6A, the lane boundary line is regularly arranged on the same traveling direction road on the same side of the center line. We use that there are many. Thereby, even if there is a noise such as a vehicle or an implantation in the road image, the presence of the lane boundary line can be estimated based on the surrounding image area. An example of the estimation result by the undetermined bright region estimation process is shown in FIG.

  Again, in FIG. 7, the number of lanes is counted based on the detected lane boundary position information for each of the roads on both sides separated by the center line of the image corresponding to the selected road vector, and the partial road lane data 26 Is stored in the storage device 20 (step 309). The partial road lane detection unit 300 performs the above-described processing of steps 301 to 310 for the corresponding images of all road vectors (step 311).

(4) Undecided Road Lane Estimation Processing Next, the undecided road lane estimation unit 400 estimates the number of lanes in the corresponding image of the road vector determined as the undecided road vector by the partial road lane detection unit 300. 13A and 13B are flowcharts showing the flow of processing by the undetermined road lane estimation unit. This process will be described with reference to examples of road vectors and vector areas of the road curve portion shown in FIG. As shown in FIG. 14, the curved road is expressed by approximating a combination of a plurality of short road vectors. Here, the road vectors Va and Vd have already determined the number of lanes, but the vectors Vb, Vc and Vx have been determined as undecided road vectors by the partial road lane detector 300 because the road vectors are shorter than the minimum length. It is. An undecided road vector is likely to be generated on a curved road in this way, but the number of lanes of the undecided road vector can also be estimated by the undecided road lane estimating unit 400. In the flowchart shown in FIG. 13, the estimation of the number of lanes of the road vector Vx is taken as an example.

  In FIG. 13A, the undetermined road lane estimating unit 400 first selects one undetermined road vector Vx (step 401), and determines the other road vectors connected from the starting point of Vx as roads for which the number of lanes has been determined. The check is performed in order until a vector is found (up to a predetermined number) (step 402). In the example shown in FIG. 14, Va corresponds to this, so the number of lanes of Va is stored in the variable Ns (step 404). Note that if no road vector for which the number of lanes has been determined is found within a predetermined number of ranges from the undetermined road vector Vx (step 403), no data is stored in Ns (step 405). Similarly, other road vectors connected from the end point of Vx are checked in order until a road vector for which the number of lanes has been determined is found (step 406). In the example shown in FIG. 14, Vd corresponds to this, so the number of lanes of Vd is stored in the variable Ne (step 408). If it is determined that no road vector for which the number of lanes has been determined is found (step 407), Ne is stored as no data (step 409).

  Subsequently, in FIG. 13B, when the number of lanes is stored in only one of Ns and Ne, the undecided road lane estimation unit 400 stores the number of lanes as the number of lanes of the road vector Vx. If neither Ns nor Ne is stored in the partial road lane data, Vx is stored in the partial road lane data of the storage device 20 as no lane number data (steps 410 to 415). . If the number of lanes is stored in both Ns and Ne, it is determined whether or not the number of lanes is the same (step 416). If the number of lanes is the same, the number of lanes is determined as the road vector Vx. Is stored in the partial road lane data of the storage device 20 (step 417). If the two data are not the same, the number of lanes of the road vector Vx is assumed to be between Ns and Ne, and is stored in the partial road lane data of the storage device 20 (step 418).

  The above processing is performed for all undecided road vectors (step 419). As described above, based on the idea that the number of lanes in the road is unlikely to change abruptly with respect to the road vector for which the number of lanes could not be determined by the partial road lane detection unit 300, The number of lanes of the undecided road vector is estimated from the number of lanes determined for the vector. FIG. 15 is a diagram showing a decision table (a) used for estimating the number of lanes of the above-mentioned undecided road vector, and an estimation result (b) of the actual number of lanes thereby.

  The number of lanes of each road in the surface image obtained through the above processing is stored as road lane data 27 of the storage device 20, and the road lane number determination program 40 of the processing device 10 ends the processing. An example of the data structure of the road lane data 27 is shown in FIG. By using the road lane data 27 together with the ground image data, a map display system, a car navigation system, a road information provision system, and the like that can provide road lane information can be configured.

  As mentioned above, although the specific embodiment was shown and demonstrated about the system and method of determining the number of road lanes of this invention, this invention is not limited to these. A person skilled in the art can make various changes and improvements to the configurations and functions of the invention according to the above-described embodiments or other embodiments without departing from the gist of the present invention.

It is a functional block diagram which shows roughly the structure of the road lane number determination system of this invention. It is a figure which shows the data structural example of the road vector data memorize | stored in the memory | storage device shown in FIG. It is a figure which shows the color model of RGB and the color model of HLS. It is a flowchart which shows the flow of a process by the search exclusion area | region detection part. It is a figure which shows the result data which detected the intersection area | region in a ground surface image. It is a figure which shows the example which detects various noise (c)-(e) from the image (a) of the road part in a ground surface image. It is a flowchart which shows the flow of the process by a partial road lane detection part. It is a figure which illustrates the brightness histogram of the row | line | column image containing a lane boundary line. It is a flowchart which shows the detail of the lane boundary determination process in step 306 of FIG. It is a flowchart which shows the detail of the undetermined bright area estimation process in step 308 of FIG. It is a figure which shows the example of the determination result obtained by the lane boundary determination process in step 306 of FIG. It is a figure which shows the example of the estimation result by the undetermined bright area estimation process in step 308 of FIG. It is a flowchart which shows the first half of the flow of the process by an undetermined road lane estimation part. It is a flowchart which shows the second half of the flow of the process by an undetermined road lane estimation part. It is a figure which shows the example of the road vector and vector area | region of a road curve part. It is a figure which shows the decision table (a) used for the lane number estimation of the undetermined road vector in an undetermined road lane estimation process, and the estimation result (b) of the actual lane number by this. It is a figure which shows the data structural example of the road lane data stored in the memory | storage device 20 shown in FIG.

Explanation of symbols

10 Processing equipment
20 Storage device
21 Ground image data
22 Road vector data
23 HLS data
24 Vegetation data
25 Search exclusion area data
26 Partial road lane data
27 Road lane data
30 I / O devices
31 Input device
32 display devices
33 Printer
40 Road Lane Number Determination Program
100 color format converter
200 Search exclusion area detector
300 Partial road lane detector
400 Undecided road lane estimation part
Vx, Vb, Vc undecided road vector
Va, Vd determined road vector
Ns Variable for storing the lane number information of the road vector on the start side of the connected road vector
Ne Variable for storing the lane number information of the road vector on the end point side of the connected road vector

Claims (10)

A system for determining the number of road lanes included in a ground image obtained by photographing a ground surface from an artificial satellite or an aircraft,
Have at least ground image data and road vector data,
Lane boundary detection means for detecting a lane boundary included in the ground image data corresponding to each road vector;
A road lane number determination system comprising: a lane number determination unit that determines the number of lanes based on the detected lane boundary line information for each road vector. A system for determining the number of road lanes included in a ground image obtained by photographing a ground surface from an artificial satellite or an aircraft,
Have at least ground image data and road vector data,
Exclusion region determination means for determining an exclusion region where lane boundary lines are not detected in the ground surface image data;
Lane boundary detection means for detecting a lane boundary included in a region other than the excluded region of the ground image data corresponding to each road vector;
A road lane number determination system comprising: a lane number determination unit that determines the number of lanes based on the detected lane boundary line information for each road vector. A system for determining the number of road lanes included in a ground image obtained by photographing a ground surface from an artificial satellite or an aircraft,
Have at least ground image data and road vector data,
Exclusion region determination means for determining an exclusion region where lane boundary lines are not detected in the ground surface image data;
A lane boundary detection means for detecting a lane boundary included in a region other than the exclusion region of the ground image data corresponding to a road vector having a predetermined length or more;
Lane number determining means for determining the number of lanes based on the information of the detected lane boundary line for each road vector;
Lane number estimating means for estimating the number of lanes for the road vector based on the number of lanes determined for one or more other road vectors in the vicinity of the road vector for the road vector not having the predetermined length or more. And a road lane number determination system. In the road lane number determination system according to any one of claims 1 to 3,
A color format conversion means for converting the ground image data between the HLS format and the RGB format;
The exclusion area determination means and the lane boundary detection means perform exclusion area determination and lane boundary detection using one or both of ground image data in HLS format and RGB format. system.   5. The road lane number determination system according to claim 4, wherein the lane boundary detection unit performs lane boundary detection using data obtained by binarizing brightness of ground image data in HLS format.   The lane boundary detection means estimates the presence of a lane boundary based on information on the lane boundary detected around the exclusion area for the exclusion area in the ground image data. The road lane number determination system according to any one of claims 3 to 5.   The lane boundary detection means divides ground image data corresponding to each road vector by a predetermined width along the road vector length direction, and determines whether each divided image data includes a lane boundary. The road lane number determination system according to any one of claims 1 to 6, wherein a lane boundary line is detected.   In the divided image data, the lane boundary detection unit is an image region having a pixel value or brightness different from other regions and having a predetermined length or more in the road vector length direction as a lane boundary line. The road lane number determination system according to claim 7, wherein the road lane number determination system is detected.   The said exclusion area | region determination means determines at least 1 among the intersection area | region, a vehicle, a vegetation, and a median strip contained in ground surface image data as an exclusion area | region, The any one of Claim 2 to 7 characterized by the above-mentioned. The road lane number determination system described.   The lane number estimating means is a road vector for which the number of lanes is determined among the other road vectors connected from the respective ends of the road vector, and for each road vector not having the predetermined length or more. Two road vectors closest to the road vector are estimated, and the number of lanes for the road vector is estimated to be a numerical value within a range where the number of lanes determined for each of the detected two road vectors is an upper limit and a lower limit. The road lane number determination system according to claim 3, wherein the road lane number determination system is a characteristic.

Images