JP4761944B2 - In-vehicle navigation device, sign recognition method, navigation system, server - Google Patents

In-vehicle navigation device, sign recognition method, navigation system, server Download PDF

Info

Publication number
JP4761944B2
JP4761944B2 JP2005334671A JP2005334671A JP4761944B2 JP 4761944 B2 JP4761944 B2 JP 4761944B2 JP 2005334671 A JP2005334671 A JP 2005334671A JP 2005334671 A JP2005334671 A JP 2005334671A JP 4761944 B2 JP4761944 B2 JP 4761944B2
Authority
JP
Japan
Prior art keywords
unit
database
information
dictionary
code
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
JP2005334671A
Other languages
Japanese (ja)
Other versions
JP2007139611A (en
Inventor
彰夫 古畑
外志正 土橋
明弘 宇田
直朗 小平
博之 水谷
智久 鈴木
Original Assignee
東芝ソリューション株式会社
株式会社東芝
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 東芝ソリューション株式会社, 株式会社東芝 filed Critical 東芝ソリューション株式会社
Priority to JP2005334671A priority Critical patent/JP4761944B2/en
Publication of JP2007139611A publication Critical patent/JP2007139611A/en
Application granted granted Critical
Publication of JP4761944B2 publication Critical patent/JP4761944B2/en
Application status is Expired - Fee Related legal-status Critical
Anticipated expiration legal-status Critical

Links

Images

Description

  The present invention relates to an in-vehicle navigation device that guides a vehicle to a destination, a sign recognition method for recognizing road markings, a navigation system, and a server.

2. Description of the Related Art Conventionally, in an in-vehicle navigation device, an image of a direction indication sign (hereinafter simply referred to as “indication”) painted on a road surface is picked up by a camera, the sign is recognized from the picked-up image, and the traveling lane of the own vehicle is determined. Judgment. Thereby, for example, it is possible to grasp whether or not the own vehicle is traveling in the correct lane before the intersection, and can instruct appropriate guidance (see, for example, Patent Document 1).
JP 2005-214883 A

  However, the sign painted on the road surface may gradually lose its shape due to aged wear or dirt, and may not be recognized correctly. In addition, the appearance of road surface images including a sign imaged by a camera is greatly different between daytime and nighttime, and stable recognition accuracy may not be obtained.

  In view of such circumstances, the present invention provides an in-vehicle navigation device, a sign recognition method, a navigation system, and a server that can recognize a sign with high accuracy even after the sign on the road surface is drowned by wear or dirt over time. It is something to be offered.

  In order to achieve the above object, an in-vehicle navigation device according to the present invention provides an imaging unit that images a road surface in front of or behind the host vehicle, a host vehicle position measurement unit that measures the position of the host vehicle, and a road marking. A dictionary unit storing corresponding marking codes and standard patterns, a dictionary comparison recognition unit for recognizing a label included in a road image captured by the imaging unit using information in the dictionary unit, position information and a marking code Information to store in association with the database unit, and to acquire the position information of the host vehicle from the host vehicle position measurement unit, and register the position information and the label code output from the dictionary comparison recognition unit in association with each other. Based on the location information of the vehicle measured by the registration unit and the vehicle position measurement unit, the database unit is searched and the corresponding label code is read, and this label code is recognized by the dictionary comparison and recognition unit. It is intended to and a database search unit for outputting as the candidate.

  According to the present invention, by narrowing down the candidate of the recognition result of the sign in the database search unit, and performing the sign recognition using the information of the dictionary part for the candidate of the recognition result of the sign, to obtain the final recognition result, The sign recognition accuracy can be improved and the calculation cost can be reduced. In particular, it is possible to narrow down the candidates for the label recognition results based on the information registered in the database part before the sign is drowned due to wear or dirt over time, so it depends on the similarity obtained by the dictionary comparison and recognition part. A high recognition result can be obtained without doing so.

  In the present invention, the information registration unit registers the registration time of the location information in the database unit in association with the location information and the labeling code, and the database search unit is the host vehicle measured by the host vehicle position measurement unit. The database section may be searched based on the position information and the current time, the corresponding label code is read, and the label code may be output to the dictionary comparison and recognition section as a recognition result candidate.

  For example, the road image captured by the imaging unit may be different between daytime and nighttime, and the appearance of the road surface image may be different due to the difference in the direction of sunlight in the morning and afternoon. However, by registering the time at the time of registration of the position information in the database unit in association with the position information and the labeling code, it is possible to register different combinations of the position information and the labeling code for each time. It is possible to reduce the variation in recognition accuracy according to.

  Furthermore, the present invention may further include a dictionary update unit that updates the dictionary unit using the labeling code obtained by the dictionary comparison and recognition unit and the feature pattern of the label that is the recognition target. .

  Thus, the sign can be recognized using the dictionary unit corresponding to the actual road marking condition, and the recognition accuracy can be improved.

  Further, the dictionary update unit newly adds a feature pattern of a label to be recognized as a standard pattern corresponding to a label code to the dictionary unit. Thereby, it is possible to accurately recognize the signs having various degrees of drowning.

  In the present invention, the information registration unit registers the similarity obtained by the dictionary comparison recognition unit in the database unit in association with the position information and the labeling code, and the database search unit performs dictionary comparison recognition for the recognition result candidate. The similarity obtained by the part is compared with the similarity registered in the database part in association with the labeling coat that is a recognition result candidate, the former similarity is lower than the latter similarity, and the difference is When the value exceeds a certain value, the dictionary comparison and recognition unit is requested to perform the label recognition by comparing the standard pattern of all the signs in the dictionary unit with the feature pattern of the label to be recognized, and through the information registration unit The database unit may be updated with the sign recognition result obtained by the dictionary comparison recognition unit. Thereby, the changed sign and the added sign can be recognized with high accuracy.

  A server based on another aspect of the present invention is a server that can be connected to one or more in-vehicle navigation devices of the present invention and the in-vehicle navigation device through a network, and associates position information with a marking code. The server database section to be stored and the database information that is a combination of the position information and the labeling code stored in the database section of the in-vehicle navigation device through the network from the in-vehicle navigation device, and compared with the information in the server database section New database information is registered in the server database unit, and the number of times the information registered in the server database unit matches the acquired database information is managed for each combination of the position information and the indication code. Location information and this based on management results Determining a combination of most probable indication code for, is to and a determination unit for transmitting the combination to the vehicle navigation apparatus via the network.

  According to the present invention, a combination of position information and a sign code is managed based on a result of integrating database information acquired from one or more in-vehicle navigation devices in a server, and a database of each in-vehicle navigation device By applying feedback to the unit, it is possible to eliminate the variation between the databases of each in-vehicle navigation device, and it is possible to improve the sign recognition accuracy in each in-vehicle navigation device.

  The server of the present invention further includes a display unit and an input unit, and the determination unit requests the in-vehicle navigation device to transmit the road surface image of the position information and the sign via the network, and responds to the request. The image information returned from the vehicle-mounted navigation device may be displayed on the display unit, and the server database unit may be updated by inputting a labeling code that is a result of the image confirmation by the operator from the input unit. The determination unit may transmit the update result of the server database unit to the vehicle-mounted navigation device through the network.

  Thereby, the reliability of the information of the database part of a server database part and a vehicle-mounted navigation apparatus can be improved, and sign recognition accuracy can be improved.

  Further, the server of the present invention generates a feature pattern of a sign included in the road image transmitted from the in-vehicle navigation device, a dictionary unit storing a sign code and a standard pattern corresponding to the sign of the road surface, It is further provided with a dictionary editing unit that newly registers a feature pattern in the dictionary unit as a standard pattern corresponding to the corresponding marking code in the dictionary unit, and a communication unit that transmits information in the dictionary unit to the vehicle-mounted navigation device via the network. Also good.

  Thereby, the recognition precision using the information of the dictionary part in a vehicle-mounted navigation apparatus can be improved.

  According to the present invention, the sign can be recognized with high accuracy even after the sign on the road surface is drowned due to aged wear or dirt.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of an in-vehicle navigation device according to an embodiment of the present invention. This in-vehicle navigation device is mounted on a vehicle such as a four-wheeled vehicle or a two-wheeled vehicle.

  In the figure, an in-vehicle navigation device 100 includes a GPS receiver 1 which is an example of a vehicle position measuring unit, an HD (Hard Disk), a CD-ROM (Compact Disc Read Only Memory), a DVD-ROM (Digital Versatile Disc). A disk drive 2 that reads map information and related information from a medium such as Read Only Memory), an operation input unit 3 that inputs a start point, a destination, etc. and various operations from a user, A display device 4 such as a CRT (cathode ray tube) or an LCD (liquid crystal display device) for displaying information on the position (own vehicle position) and its surrounding map information and navigation of the own vehicle, and audio information related to navigation are output. The speaker 5 is configured to include a camera 6 that is an imaging unit that images a road surface in front of or behind the host vehicle, and a control device 7.

  The control device 7 stores a CPU (Central Processing Unit) 7a, a RAM (Random Access Memory) 7b that provides a work area and a data storage area for the CPU 7a, a program executed by the CPU 7a, various data, and the like. ROM (Random Only Memory) 7c and a flash memory 7d which is an electrically erasable and rewritable ROM.

  The GPS receiver 1 measures the current position of the own vehicle by GPS (Global Positioning System), detects the current position of the own vehicle (own vehicle position), and controls the detection result of the own vehicle. 7 is provided.

  The camera 6 is mounted on the front part or rear part of the body of the host vehicle, captures an image of the road ahead or behind the host vehicle, and captures an image of a sign such as straight, left turn, or right turn painted on the road surface. It is means of. Specifically, the camera 6 is an image sensor such as a charge-coupled device (CCD), for example, and a signal processing circuit that converts a signal imaged by the image sensor from an analog signal to a digital signal and outputs the signal to the control device 7. And have.

  FIG. 2 is a block diagram showing the configuration of the sign recognition system incorporated in the control device 7. The sign recognition refers to generating a code (marking code) corresponding to the sign from the image of the sign on the road surface. Based on this marking code, the traveling lane of the host vehicle on a road having a plurality of lanes can be determined, and the determination result can be used for navigation (route guidance) of the host vehicle.

  As shown in FIG. 2, the sign recognition system includes an image processing unit 11, a dictionary unit 12, a sign recognition unit 13, an information registration unit 14, a database unit 15, a clock unit 16, and the like.

  The image processing unit 11 is a unit that performs processing such as noise removal, perspective transformation, and binarization on the image data captured by the camera 6 and passes the result to the sign recognition unit 13.

  The dictionary unit 12 is a dictionary in which information of standard patterns of various signs used in the sign recognition is stored in association with the sign codes. The dictionary unit 12 is stored in, for example, the flash memory 7d in the control device 7 so that it can be updated as appropriate. Alternatively, if an HDD is mounted as the disk drive 2, the dictionary unit 12 may be stored in the HDD, loaded from the HDD into the RAM 7b of the control device 7 during operation, and accessed from the CPU 7a. .

  As shown in FIG. 3, the database unit 15 is a part that stores information such as a labeling code, similarity, position, and time in association with each other. The database unit 15 is stored in, for example, a flash memory 7d in the control device 7. Alternatively, if an HDD is mounted as the disk drive 2, the database unit 15 is stored in the HDD, and loaded into the RAM 7 b of the control device 7 from the HDD during operation so that the CPU 7 a can access it at high speed. Good.

  The sign recognition unit 13 includes a cutout unit 13b, a database search unit 13a, a feature extraction unit 13c, and a dictionary comparison recognition unit 13d.

  The database search unit 13a is a unit that searches the database unit 15 based on the information on the position of the vehicle acquired from the GPS receiver 1, the time information acquired from the clock unit 16, and the like and determines a candidate for the recognition result of the indication. is there. The cutout unit 13b is means for cutting out a marking portion from the binary pattern information of the road surface image. The feature extraction unit 13c is a means for generating feature pattern information by extracting the features of the binary pattern of the marking portion cut out by the cutout unit 13b. The dictionary comparison / recognition unit 13d includes a standard pattern of a label determined as a candidate for a label recognition result by the database search unit 13a among the standard patterns of the label stored in the dictionary unit 12, and a feature generated by the feature extraction unit 13c. This is means for comparing a pattern with, for example, a pattern matching method and outputting a label code associated with the most similar standard pattern as a recognition result.

  The clock unit 16 is means for outputting information on the current time. Each time the sign recognition unit 13 obtains a recognition result of the sign, the information registration unit 14 displays the sign code and similarity information as the recognition result, and the vehicle position acquired by the sign recognition unit 13 from the GPS receiver 1. And the time information acquired from the clock unit 16 are associated with each other and registered in the database unit 15.

  Next, the operation of the in-vehicle navigation device 100 of this embodiment will be described.

  FIG. 4 is a flowchart showing a procedure for recognizing road markings in the in-vehicle navigation device 100 of the present embodiment.

  During operation of the in-vehicle navigation device 100 of the present embodiment, an image obtained by imaging the road surface in front of or behind the host vehicle is captured by the camera 6 mounted on the front portion of the vehicle (step S401). FIG. 5 is a diagram illustrating an example of a captured image (digital image) of the camera 6. In this example, as a photographed image of the camera 6, for example, a road surface image is obtained in which a sign 23 indicating straight travel is painted between the center line 21 and the road line 22.

  The road surface image captured by the camera 6 is input to the control device 7. The control device 7 first performs correction such as removing a high-band signal as noise in the image processing unit 11 with respect to the input road surface image, and then performs perspective transformation (step S402). The perspective transformation is, for example, a process of coordinate-transforming a road surface image from an oblique viewpoint as shown in FIG. 5 into a road image from an overhead viewpoint as shown in FIG. However, in this perspective transformation, the standard pattern of the marking in the dictionary unit 12 used in the subsequent sign recognition is a pattern from an oblique viewpoint, and the sign recognition is correctly performed on a road surface image from the oblique viewpoint. This is not necessary if the system is such that

  Next, the image processing unit 11 performs binarization processing on the perspective-transformed image to obtain binary pattern information of the road surface image from which components unnecessary for recognition are removed (step S403). Subsequently, the binary pattern information of the road surface image is sent from the image processing unit 11 to the cutout unit 13b of the sign recognition unit 13, and a trigger signal is sent to the database search unit 13a of the sign recognition unit 13. When receiving the trigger signal, the database search unit 13a of the sign recognition unit 13 takes in the information on the vehicle position from the GPS receiver 1 and takes in the time information from the clock unit 16, and uses the database unit 15 based on these information. Search is performed (step S404).

  If there is corresponding information (record) by this search (YES in step S405), the labeling code in the corresponding information is output to the dictionary comparison and recognition unit 13d as a candidate for the label recognition result (step 406). . If there is no corresponding information in the database unit 15 (NO in step S405), the dictionary comparison recognition unit 13d is notified that there is no candidate (step S407). Here, the database search unit 13a of the sign recognition unit 13 may search the database unit 15 based only on the information on the vehicle position acquired from the GPS receiver 1 to determine a candidate for the recognition result of the sign. Good. The determination of the candidate for the recognition result of the sign by searching the database unit 15 will be described in detail later.

  On the other hand, when the cutout unit 13b of the sign recognition unit 13 receives the binary pattern information of the road surface image from the image processing unit 11, the cutout unit 13b performs a process of cutting out the mark part from the binary pattern information of the road surface image (step S408). The marking portion is cut out, for example, by performing a projection process in the lane direction and the lane width direction on the binary pattern information of the road surface image, and based on the result of the projection process. FIG. 7 shows the result of the projection processing for the binary pattern information of the road surface image. In the result of the projection processing in the lane direction shown in the figure, peaks P21, P23, and P22 stand at three places corresponding to the center line 21, the marking 23, and the road line 22 on the road surface. The cutout unit 13b detects these three peaks P21, P23, and P22 by comparison with a predetermined threshold, and determines that the position of the middle peak P23 is a marking portion in the lane width direction. Further, the cutout unit 13b determines the position of the marking portion in the lane direction by performing projection processing in the lane width direction in the same manner. The cutout unit 13b can cut out the pattern information 24 of the marking portion from the binary pattern information of the road surface image based on the respective determination results.

  Next, the feature extraction unit 13c of the sign recognition unit 13 extracts features such as density, contour, and shape from the extracted pattern information of the sign part, and the resulting feature pattern is stored in the dictionary comparison recognition unit 13d. Output (step S409).

  When the dictionary comparison and recognition unit 13d acquires a candidate for the label recognition result from the database search unit 13a, the standard pattern for the label that is a candidate for the label recognition result among the standard patterns for all the labels stored in the dictionary unit 12 And the feature pattern input from the feature extraction unit 13c are compared with each other by, for example, a pattern matching method, and the respective similarities are obtained, and the labeling code associated with the standard pattern with the highest similarity is output as the recognition result. (Step S410a). When the dictionary comparison recognition unit 13d receives a notification indicating that there is no candidate for the label recognition result from the database search unit 13a, the dictionary comparison recognition unit 13d generates the standard pattern and feature extraction unit 13c for all the signs stored in the dictionary unit 12. Compared with the feature pattern of the sign, the sign code associated with the standard pattern having the highest similarity is output as a recognition result (step S410b).

  Next, the sign recognition unit 13 outputs the sign code and the similarity, which are the recognition results obtained by the dictionary comparison recognition unit 13d, to the information registration unit 14, and when the database search unit 13a receives the trigger signal. Information on the own vehicle position acquired from the GPS receiver 1 and time information acquired from the clock unit 16 are output to the information registration unit 14, respectively. Upon receiving these pieces of information, the information registration unit 14 registers these pieces of information in association with each other in the database unit 15 (step S411).

  By repeating the above procedure, information related to road markings existing on the route through which the host vehicle has passed is accumulated in the database unit 15. In the lane before the intersection, a plurality of signs such as left turn, straight ahead, and right turn are painted side by side, and these are regarded as signs at the same position in the vehicle-mounted navigation device 100. Therefore, a plurality of records with different labeling codes associated with the same position information are recorded in the database unit 15.

  By the way, the position information registered in the database unit 15 has some variations even if the position information is actually the same position due to a positioning error of GPS. Therefore, the position information acquired from the GPS receiver 1 is compared with the position information on the map information, and if the difference is less than a certain value, the position information on the database unit 15 is replaced with the position information on the map information. It is preferable to register and search for information.

  It is desirable that the time information acquired from the clock unit 16 is converted into information obtained by dividing a day into several time zones, for example, daytime and nighttime, and registration and search in the database unit 15 are performed. . During the day, it may be further divided into afternoon and morning. The road surface image captured by the camera 6 has different aspects during the day and at night, and the aspect of the road surface image may also differ due to the difference in the direction of sunlight between morning and afternoon. For this reason, managing information on the sign for each time zone leads to improvement in recognition accuracy.

  Further, similarity information is also registered in the database unit 15, but for the similarity, only the maximum value obtained in the past is registered in the database unit 15 for each different combination of position information and time information. It is like that. Further, a minimum similarity value required for registration in the database unit 15 may be determined, and registration in the database unit 15 may be invalidated when the value is not less than this value.

  Next, the operation of searching the database unit 15 based on the information on the vehicle position and time and determining a candidate for the recognition result of the sign will be described. FIG. 8 is a flowchart showing an operation for determining a candidate for the sign recognition result.

  First, the database search unit 13a of the sign recognition unit 13 acquires information on the vehicle position detected by the GPS receiver 1, and also acquires time information from the clock unit 16 (step S801). It is determined whether information (record) whose position and position information match is present in the database unit 15 (step S802). If there is no matching information in the database unit 15 (NO in step S802), it is determined that there is no candidate for the label recognition result, and the notification is notified to the dictionary comparison recognition unit 13d (step S805). ). If information matching the vehicle position exists in the database unit 15 (YES in step S802), the database search unit 13a includes the time acquired from the clock unit 16 among the one or more pieces of information matched in position information. It is determined whether there is information that matches the time information (step S803). If there is no information matching the time information in the database unit 15 (NO in step S803), it is determined that there is no candidate for the label recognition result, and the notification is notified to the dictionary comparison recognition unit 13d (step S805). ). If there is information that matches the time information in the database unit 15 (YES in step S803), the database search unit 13a determines the labeling code included in the information as a candidate for the label recognition result, and sends it to the dictionary comparison recognition unit 13d. Output (step S804).

  In addition, when the information with which time information corresponds is not obtained from the database part 15, the information obtained by the search of the database part 15 based on position information is output to the dictionary comparison recognition part 13d as a candidate of a label recognition result. You may do it.

  As described above, according to the present embodiment, candidates for the sign recognition result are narrowed down by the database search unit 13a, and the standard pattern of the dictionary part 12 and the feature pattern of the sign are compared for the sign recognition result candidate. By obtaining the sign recognition result, the recognition accuracy can be improved and the calculation cost can be reduced.

  Here, a specific case in which the sign recognition method of the present embodiment can particularly effectively work will be described with respect to improvement in recognition accuracy.

  FIG. 9 is an example of an image captured by the camera 6. The photographed road surface is painted with a sign 23 indicating straight travel between the center line 21 and the road side line 22, but a part of the sign 23 is scraped by repeated contact with the vehicle tire. Yes. FIG. 10 shows the result of performing perspective transformation on this captured image, and FIG. 11 shows the result of determining the cutout position 24 of the marking portion by sequentially performing binarization processing and projection processing on the perspective transformed image of FIG. Show. In such a case, the dictionary comparison / recognition unit 13d cannot obtain a high similarity with the standard pattern of any sign stored in the dictionary unit 12, and it is difficult to obtain a correct recognition result. In the present embodiment, candidates for the label recognition result can be narrowed down based on information registered in the database unit 15 before the label is drowned due to wear or dirt over time. A high recognition result can be obtained without depending on the degree of similarity.

  Further, in the in-vehicle navigation device 100 according to the present embodiment, the dictionary unit 12 is stored in a rewritable storage area, so the contents of the dictionary unit 12 are used by using a labeling code and an input feature pattern as a recognition result. You may make it update. The updating of the dictionary unit 12 may be automatically performed when, for example, the similarity obtained by the dictionary comparison recognition unit 13d is less than a certain value. The updating of the dictionary unit 12 here means adding a standard pattern of marking separately from the original one.

(Modification of step S410a)
Next, a modification of step S410a in FIG. 4 will be described. Step S410a in FIG. 4 compares the standard pattern of the label of the candidate of the label recognition result with the feature pattern of the label to obtain the respective similarity, and the labeling code associated with the standard pattern with the highest similarity Is output as a recognition result. However, when a change or addition is made to the actual road surface marking, the similarity between the standard pattern of the marking of the candidate of the marking recognition result and the input feature pattern is significantly lower than the similarity registered in the database unit 15. It may be possible.

  Therefore, in this modification, the following functions are incorporated in the database search unit 13a. That is, the database search unit 13a uses the maximum similarity obtained by recognizing the candidate of the label recognition result determined by itself in the dictionary comparison recognition unit 13d and the sign coat that is a candidate of the label recognition result. The similarities registered in the database unit 15 in association with each other are compared. By this comparison, the database search unit 13a, when the similarity of the former is lower than the similarity of the latter and the difference exceeds a certain value, all the indications of the dictionary unit 12 are displayed to the dictionary comparison recognition unit 13d. A request is made to recognize the sign by comparing the standard pattern with the feature pattern of the sign that is the recognition target. In response to this request, the dictionary comparison / recognition unit 13d compares the standard pattern of all the signs in the dictionary unit 12 with the feature pattern of the sign that is the recognition target, and obtains the sign recognition result again. The database search unit 13a requests the information registration unit 14 to update the database unit 15 with the sign recognition result newly obtained by the dictionary comparison recognition unit 13d, and the information registration unit 14 follows the sign recognition result according to this request. The database unit 15 is updated. Thereby, it becomes possible to cope with a change or addition of the road marking.

  Next, a second embodiment of the present invention will be described.

FIG. 12 is a diagram showing a configuration of a navigation system according to the second embodiment.
As shown in the figure, the navigation system 200 is composed of one or more in-vehicle navigation devices 110,... And a server 150. The in-vehicle navigation device 110,... And the server 150 are connected through a network 180 such as the Internet.

  The in-vehicle navigation device 110 is obtained by adding a communication unit 8 for communicating with the server 150 through the network 180 to the configuration of the in-vehicle navigation device 100 of the first embodiment. In addition, the control device 7 of the vehicle-mounted navigation device 110 controls the communication unit 8 to periodically transmit information of the database unit 15 to the server 150 or to display a road marking at a position specified by the server 150. An image or the like is transmitted, and update data of the database unit 15 and update data of the dictionary unit 12 are received from the server 150 through the communication unit 8 to update the database unit 15 and update the dictionary unit 12. It is possible to control.

  On the other hand, as shown in FIG. 13, the server 150 includes a control device 151, a communication unit 152, a display unit 153, an input unit 154, a disk drive 155, and the like. Specifically, the control device 151 includes a CPU, a RAM for providing a work area and a data storage area for the CPU, and a ROM in which programs executed by the CPU and various parameters are fixedly stored. It is configured. The communication unit 152 is means for communicating with each of the in-vehicle navigation devices 110 through the network 180. The display unit 153 is a display device such as a CRT or LCD that displays a captured image of the camera 6 transferred from the in-vehicle navigation device 110. The input unit 154 is an input device such as a keyboard and a mouse for receiving input from the operator of the server 150. The disk drive 155 is a recording device that can read and write data and programs with respect to media such as an HDD.

  The control device 151 includes a determination unit 161, a dictionary editing unit 162, a dictionary unit 163, a database unit 164 that is a server database unit, and the like.

  The dictionary unit 163 is a dictionary in which information of standard patterns of various signs used in the sign recognition is stored in association with the sign codes. This dictionary unit 163 is stored in, for example, the disk drive 155 so that it can be updated as appropriate, and is loaded from the disk drive 155 into the RAM of the control device 151 and accessed from the CPU during operation.

  The dictionary editing unit 162 generates a feature pattern by extracting features of the image information (for example, a cutout image of the marking portion) received from the in-vehicle navigation device 110, and the feature pattern and the standard of the marking stored in the dictionary unit 163 The similarity is calculated by comparing the pattern with, for example, a pattern matching method. If the calculated similarity is less than a predetermined value, the dictionary editing unit 162 uses the generated feature pattern as an original standard pattern as a standard pattern corresponding to the corresponding label code in the dictionary unit 163. Separately, it is newly registered in the dictionary unit 163.

  The database unit 164 is a storage unit that stores combinations of marking codes and position information. The database unit 164 is accessed from the CPU, for example, by storing a backup of the database unit 164 in the disk drive 155 and loading it from the disk drive 155 to the RAM of the control device 151 during operation.

  The determination unit 161 receives a combination of the indication code and the position information, which is database information, from the in-vehicle navigation device 110 via the network 180 in the communication unit 152 and compares it with the contents of the database unit 164 of the server 150. It is means for controlling the communication unit 152 so as to update the database unit 164 and to transmit information in the database unit 164 of the server 150 to the in-vehicle navigation device.

  Next, the operation of the navigation system 200 will be described.

  FIG. 14 is a diagram showing a processing sequence relating to the operation of the navigation system 200.

  The control device 7 of the in-vehicle navigation device 110 controls the communication unit 8 and periodically transmits all combinations of the labeling code and the position information in the database unit 15 to the server 150 through the network 180 as database information. It transmits to the server 150 (step S1401).

  The server 150 receives the combination of the indication code and the position information, which is database information, from the in-vehicle navigation device 110 via the network 180 in the communication unit 152 (step S1402), and the individual database information acquired from the in-vehicle navigation device 110. The (combination of one position information and a labeling code) is compared with the database unit 164 of the server 150 (step S1403).

  As a result of this comparison, when position information that matches the database unit 164 of the server 150 is not registered (NO in step S1404), the determination unit 161 determines the database information acquired from the in-vehicle navigation device 110 as new information. Then, it is registered in the database unit 164 of the server 150 (step S1405). If the determination unit 161 determines that the corresponding position information is registered in the database unit 164 of the server 150 (YES in step S1404), the determination unit 161 associates the position information with the database unit 164 of the server 150. The registered code and the labeled code in the database information acquired from the in-vehicle navigation device 110 are compared to determine whether there is a matching labeled code (step S1406). There may be a plurality of label codes registered in association with one piece of position information. When the labeling code in the database information acquired from the vehicle-mounted navigation device 110 does not match any labeling code registered in the database unit 164 of the server 150 (NO in step S1406), the determining unit 161 The database information acquired from the navigation device 110 is determined as new information and registered in the database unit 164 of the server 150 (step S1405).

  Further, if the labeling code in the database information acquired from the in-vehicle navigation device 110 matches any of the marking codes registered in the database unit 164 of the server 150 (YES in step S1406), the position information and It is determined whether or not a flag indicating these probabilities is set in the database unit 164 for the labeling code (step S1407). Here, the flag indicating the probability is, for example, for a label code having the highest number of past hits among one or more label codes registered in the database unit 164 of the server 150 in association with one position information. Only to be set.

  If the determination unit 161 determines that the flag is set (YES in step S1407), the determination unit 161 increments the position information and the number of marking code hits in which the flag is set (step S1408). In this case, the process ends here. When the flag is not set (NO in step S1407), the determination unit 161 increments the position information and the number of indication code hits acquired from the in-vehicle navigation device 110 (step S1409), and the position where the flag is set. The combination of the information and the marking code is held in the RAM as information to be transmitted to the vehicle-mounted navigation device 110 (step S1410).

  The determination unit 161 repeats the same process for all database information acquired from the in-vehicle navigation device 110, and when the process for all database information is completed (YES in step S1411), the position information stored in the RAM and The communication unit 152 is controlled so that information on the combination with the labeling code is collected and transmitted to the vehicle-mounted navigation device 110 (step S1412).

  When the in-vehicle navigation device 110 receives the combination of the position information and the sign code from the server 150 (step S1413), the in-vehicle navigation apparatus 110 acquires the sign code registered in association with the corresponding position information in the database unit 15 from the server 150. The updated sign code is updated (step S1414).

  As described above, according to the present embodiment, the server 150 manages the combination of the position information and the labeling code based on the result of integrating the database information acquired from the plurality of in-vehicle navigation devices 110, By applying feedback to the database unit 15 of each in-vehicle navigation device 110, it is possible to eliminate the variation between the databases in each in-vehicle navigation device 110, and to improve the sign recognition accuracy in each in-vehicle navigation device 110. Can do.

  In addition, the determination unit 161 of the server 150 monitors the difference in the number of hits for a plurality of marking codes associated with the same position information, and after the total number of hits reaches a required value, If the hit difference is within a certain value, the vehicle-mounted navigation device 110 is requested to transmit an image including position information. When the determination unit 161 of the server 150 receives the image information transmitted from the in-vehicle navigation device 110 in response to this request (for example, a cutout image of the marking portion), the determination unit 161 displays the image information on the display unit 153 and Control is performed so as to prompt visual confirmation by the operator. The determination unit 161 receives the result of confirmation from the operator, that is, the labeling code and the position information from the input unit 154, and updates the information in the database unit 164. Thereby, the reliability of the information of the database part 164 can be improved. The updated information in the database unit 164 is transmitted to the in-vehicle navigation device 110, and the database unit 15 in the in-vehicle navigation device 110 is updated.

  Further, the determination unit 161 passes the image received from the in-vehicle navigation device 110 to the dictionary editing unit 162. The dictionary editing unit 162 generates a feature pattern by extracting features of the image information (for example, a cutout image of the marking portion) received from the in-vehicle navigation device 110, and the feature pattern and the standard of the marking stored in the dictionary unit 163 The similarity is calculated by comparing the pattern with, for example, a pattern matching method. If the calculated similarity is less than a predetermined value, the dictionary editing unit 162 uses the generated feature pattern as a standard pattern corresponding to the corresponding labeling code in the dictionary unit 163 separately from the original standard pattern. New registration in the part 163. The updated information of the dictionary unit 163 is transmitted to the in-vehicle navigation device 110 and is replaced with the dictionary unit 12 of the in-vehicle navigation device 110. Thereby, the recognition accuracy of the sign in the vehicle-mounted navigation device 110 can be improved.

  When the dictionary editing unit 162 generates a feature pattern without comparing the generated feature pattern with the standard pattern of the dictionary unit 163 and calculating the similarity, the dictionary editing unit 162 unconditionally corresponds to the label code. You can also add additional registration.

  In the embodiment described above, the in-vehicle navigation device uses only GPS to measure the current position of the host vehicle. However, more accurate from the output of the gyro sensor and vehicle speed pulse mounted on the vehicle and the GPS signal. You may make it measure a self-vehicle position.

  The present invention is not limited only to the above-described embodiment, and it is needless to say that various updates can be added without departing from the gist of the present invention.

It is a block diagram which shows the structure of the vehicle-mounted navigation apparatus concerning one Embodiment of this invention. It is a block diagram which shows the structure of the mechanism regarding the label | marker recognition in the control apparatus in FIG. It is a figure which shows the structure of the database in FIG. It is a flowchart which shows the procedure of the road marking recognition in the vehicle-mounted navigation apparatus of FIG. It is a figure which shows the example of the picked-up image of a camera. It is a figure which shows the perspective transformation result of the picked-up image of FIG. It is a figure which shows the result of the projection process with respect to the binary pattern information of the road surface image of FIG. It is a flowchart which shows the determination operation | movement of the candidate of a sign recognition result. It is a figure which shows the example of the picked-up image of a camera. It is a figure which shows the result of having performed perspective transformation with respect to the picked-up image of FIG. It is a figure which shows extraction of the marking part from the picked-up image of FIG. It is a figure which shows the structure of the navigation system concerning 2nd Embodiment. It is a figure which shows the structure of the server in the navigation system of FIG. It is a figure which shows the process sequence regarding operation | movement of the navigation system of FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... GPS receiver, 6 ... Camera, 7 ... Control apparatus, 8 ... Communication part, 11 ... Image processing part, 12 ... Dictionary part, 13 ... Marking recognition part, 13a ... Database search part, 13b ... Extraction part, 13c ... Feature extraction unit, 13d ... Dictionary comparison recognition unit, 14 ... Information registration unit, 15 ... Database unit, 16 ... Clock unit, 100, 110 ... In-vehicle navigation device, 150 ... Server, 151 ... Control device, 152 ... Communication unit, 153 ... Display unit, 154 ... Input unit, 155 ... Disk drive, 161 ... Determination unit, 162 ... Dictionary editing unit, 163 ... Dictionary unit, 164 ... Database unit, 180 ... Network, 200 ... Navigation system.

Claims (10)

  1. An imaging unit that captures a road surface in front of or behind the host vehicle and acquires a road surface image ;
    A vehicle position measuring unit that measures the position of the vehicle and outputs vehicle position information ;
    A dictionary section storing a marking code and a standard pattern corresponding to the marking on the road lane on the road surface ;
    A database unit for storing the labeling code corresponding to the marking on the road lane on the road surface and its position information;
    A database search unit that searches for candidates for markings on the travel lane of the host vehicle from the database unit using the host vehicle position information, and outputs a corresponding marking code;
    Of the marking codes stored in the dictionary unit, the road pattern image is compared with a standard pattern corresponding to the marking code output by the database search unit, and the marking on the traveling lane of the host vehicle is acquired and the corresponding marking is obtained. A dictionary comparison and recognition unit that outputs codes ;
    An information registration unit that registers the vehicle position information and the label code output by the dictionary comparison recognition unit in association with each other in the database unit;
    A dictionary updating unit that updates the standard pattern of the dictionary unit using the marking code output by the dictionary comparison recognition unit and the feature pattern of the marking on the traveling lane that appears in the road surface image. A vehicle-mounted navigation device.
  2. The information registration unit registers the registration time of the vehicle position information in the database unit in association with the vehicle position information and the label code output by the dictionary comparison recognition unit ,
    The database search unit searches for a candidate for a label on a travel lane of the host vehicle from the database unit using the own vehicle position information and the current time, and outputs a corresponding label code. The in-vehicle navigation device according to Item 1.
  3. It said dictionary updating unit, the dictionary unit, vehicle navigation apparatus according to claim 1, characterized by adding the feature pattern as a standard pattern corresponding to the sign code the dictionary comparing recognition unit has output.
  4. The information registration unit registers the similarity obtained by the comparison in the dictionary comparison recognition unit in the database unit in association with the vehicle position information and the label code output by the dictionary comparison recognition unit ,
    Wherein the database search unit, the search for the results of the candidates and the similarity obtained by the comparison in the dictionary comparing recognition unit, the search result candidate and is marked in association with the code similar registered in the database unit When the similarity of the former is lower than the similarity of the latter and the difference exceeds a certain value, the standard of all the signs stored in the dictionary unit for the dictionary comparison recognition unit was asked to compare the characteristic patterns of markings on the traveling lane appearing in pattern and the road image, through the information registering unit, updating the database unit with a comparison result in said dictionary comparing recognition unit The in-vehicle navigation device according to claim 1.
  5. Capture the road surface in front of or behind the host vehicle with the imaging unit to obtain a road image,
    By measuring the position of the host vehicle by a position measurement unit to acquire host vehicle position information,
    A candidate for a sign on the travel lane of the host vehicle from a database part that stores the label code and the position information corresponding to the sign on the travel lane on the road surface using the own vehicle position information by the database search unit. And output the corresponding label code
    The standard pattern corresponding to the marking code output by the database search unit among the marking codes stored in the dictionary unit storing the marking code and the standard pattern corresponding to the marking on the road lane on the road surface by the dictionary comparison recognition unit And the road surface image, obtain a sign on the traveling lane of the host vehicle and output a corresponding sign code,
    The dictionary updating unit updates the standard pattern of the dictionary unit using the marking code output from the dictionary comparison and recognition unit and the feature pattern of the marking on the traveling lane appearing on the road surface image. A sign recognition method characterized by
  6. A navigation system comprising one or more in-vehicle navigation devices according to claim 1 and a server connectable to the in-vehicle navigation device through a network,
    The in-vehicle navigation device is
    Means for transmitting, as database information, a combination of position information and marking code stored in the database unit to the server through the network;
    The server
    A server database unit that stores the positional information and the labeling code in association with each other;
    The database information is acquired from the in-vehicle navigation device through the network, compared with the information in the server database unit, new database information is registered in the server database unit, and information registered in the server database unit And the number of times the acquired database information is matched for each combination of the position information and the labeling code, and based on this management result, the combination of the position information and the most probable labeling code is determined. A navigation system comprising: a determination unit that determines and transmits the combination to the in-vehicle navigation device through the network.
  7. One or more in-vehicle navigation devices according to claim 1 and a server connectable to the in-vehicle navigation device through a network,
    A server database unit that stores the positional information and the labeling code in association with each other;
    The database information that is a combination of the position information and the labeling code stored in the database unit of the in-vehicle navigation device is acquired from the in-vehicle navigation device through the network, and compared with the information in the server database unit, The database information is registered in the server database unit, and the number of times the information registered in the server database unit matches the acquired database information is managed for each combination of the position information and the indication code, A determination unit that determines a combination of position information and the most probable labeling code based on the management result, and transmits the combination to the in-vehicle navigation device through the network, Server.
  8. A display unit and an input unit;
    The determination unit
    The vehicle navigation device is requested to transmit the road surface image with the position information and the sign through the network, the image information responded from the vehicle navigation device in response to the request is displayed on the display unit, and the operator The server according to claim 7 , wherein the server database unit is updated by inputting a labeling code, which is an image confirmation result, from the input unit.
  9. The server according to claim 8 , wherein the determination unit transmits an update result of the server database unit to the in-vehicle navigation device through the network.
  10. A dictionary portion storing a marking code and a standard pattern corresponding to the marking on the road lane of the road surface;
    Dictionary editing for generating a feature pattern of a sign appearing in a road image transmitted from the in-vehicle navigation device and newly registering the feature pattern in the dictionary unit as a standard pattern corresponding to a corresponding sign code of the dictionary unit And
    The server according to claim 9 , further comprising a communication unit that transmits information of the dictionary unit to the in-vehicle navigation device through the network.
JP2005334671A 2005-11-18 2005-11-18 In-vehicle navigation device, sign recognition method, navigation system, server Expired - Fee Related JP4761944B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2005334671A JP4761944B2 (en) 2005-11-18 2005-11-18 In-vehicle navigation device, sign recognition method, navigation system, server

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2005334671A JP4761944B2 (en) 2005-11-18 2005-11-18 In-vehicle navigation device, sign recognition method, navigation system, server

Publications (2)

Publication Number Publication Date
JP2007139611A JP2007139611A (en) 2007-06-07
JP4761944B2 true JP4761944B2 (en) 2011-08-31

Family

ID=38202666

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2005334671A Expired - Fee Related JP4761944B2 (en) 2005-11-18 2005-11-18 In-vehicle navigation device, sign recognition method, navigation system, server

Country Status (1)

Country Link
JP (1) JP4761944B2 (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4948267B2 (en) 2007-05-25 2012-06-06 東芝機械株式会社 Sheet / film forming apparatus and sheet / film forming method
DE102007030226A1 (en) 2007-06-29 2009-01-08 Robert Bosch Gmbh Camera-based navigation system and method for its operation
JP2009180631A (en) * 2008-01-31 2009-08-13 Denso It Laboratory Inc Navigator, navigation method and program
JP4940168B2 (en) * 2008-02-26 2012-05-30 日立オートモティブシステムズ株式会社 Parking space recognition device
JP2012216102A (en) * 2011-04-01 2012-11-08 Hitachi Kokusai Electric Inc Image storage system
JP6431261B2 (en) * 2013-12-20 2018-11-28 矢崎エナジーシステム株式会社 Operation information management system

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001249686A (en) * 2000-03-08 2001-09-14 Matsushita Electric Ind Co Ltd Method and device for recognizing speech and navigation device
JP2002243469A (en) * 2001-02-19 2002-08-28 Alpine Electronics Inc Navigation device and map database updating system
JP4581290B2 (en) * 2001-05-16 2010-11-17 パナソニック株式会社 Speech recognition apparatus and speech recognition method
JP3817631B2 (en) * 2001-12-10 2006-09-06 アルパイン株式会社 Road warning detection device
JP2004164466A (en) * 2002-11-15 2004-06-10 Sony Corp Information updating system, information processor, and information updating method
JP4218441B2 (en) * 2003-06-27 2009-02-04 株式会社デンソー On-vehicle moving body detection device

Also Published As

Publication number Publication date
JP2007139611A (en) 2007-06-07

Similar Documents

Publication Publication Date Title
JP3496479B2 (en) Road data maintenance system
US8359156B2 (en) Map generation system and map generation method by using GPS tracks
EP1975565B1 (en) Road surface feature information collecting apparatus and method
US8452101B2 (en) Feature extraction method, and image recognition method and feature database creation method using the same
KR101241651B1 (en) Image recognizing apparatus and method, and position determining apparatus, vehicle controlling apparatus and navigation apparatus using the image recognizing apparatus or method
JP4861850B2 (en) Lane determination device and lane determination method
US8174568B2 (en) Unified framework for precise vision-aided navigation
JP4654823B2 (en) Road map data update system and road detection system
JP2005265494A (en) Car location estimation system and drive support device using car location estimation system and drive support device using this
DE102009054286B4 (en) Character recognition device
US20130057686A1 (en) Crowd sourcing parking management using vehicles as mobile sensors
US8284995B2 (en) Method for updating a geographic database for an in-vehicle navigation system
US20120288138A1 (en) System and method for traffic signal detection
US8351704B2 (en) Method of capturing linear features along a reference-line across a surface for use in a map database
US20080243378A1 (en) System and method for vehicle navigation and piloting including absolute and relative coordinates
JP2009014574A (en) Road information generator, road information generation method, and road information generation program
CN102565832B (en) Method of augmenting GPS or gps/sensor vehicle positioning using additional in-vehicle vision sensors
US20080218383A1 (en) System and method for parking infraction detection
JP2007108043A (en) Location positioning device, location positioning method
US8379923B2 (en) Image recognition processing device, method, and program for processing of image information obtained by imaging the surrounding area of a vehicle
JP4437556B2 (en) Feature information collecting apparatus and feature information collecting method
CN101675442B (en) Lane determining device, lane determining method and navigation apparatus using the same
US7756617B1 (en) Vehicular monitoring system
JP6325806B2 (en) Vehicle position estimation system
EP1686538A2 (en) Vehicle position recognizing device and vehicle position recognizing method

Legal Events

Date Code Title Description
A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20081017

A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20110121

A131 Notification of reasons for refusal

Free format text: JAPANESE INTERMEDIATE CODE: A131

Effective date: 20110125

A521 Written amendment

Free format text: JAPANESE INTERMEDIATE CODE: A523

Effective date: 20110328

RD02 Notification of acceptance of power of attorney

Free format text: JAPANESE INTERMEDIATE CODE: A7422

Effective date: 20110328

A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20110510

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20110607

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20140617

Year of fee payment: 3

R150 Certificate of patent or registration of utility model

Free format text: JAPANESE INTERMEDIATE CODE: R150

S531 Written request for registration of change of domicile

Free format text: JAPANESE INTERMEDIATE CODE: R313531

R350 Written notification of registration of transfer

Free format text: JAPANESE INTERMEDIATE CODE: R350

LAPS Cancellation because of no payment of annual fees