Information processing apparatus and information processing method

Abstract

The present invention provides an information processing apparatus and an information processing method capable of appropriately processing a blocked area of a road. The information processing apparatus of the present invention includes a first information conversion section (15) that infers a first road area from an image captured of a road, a second information conversion section (17) that infers a second road area from map information stored in a map information storage section (11), and a difference extraction section (16) that extracts difference information indicating a difference between the first road area and the second road area. In the information processing method of the present invention, a first road area is inferred from an image captured of a road by a computer, a second road area is inferred from map information, and difference information indicating a difference between the first road area and the second road area is extracted.

Description

TECHNICAL FIELD

[0001] The present application relates to an information processing apparatus and an information processing method. For example, an information processing apparatus and an information processing method that processes map data for an automated driving or advanced safety driving system. BACKGROUND

[0002] For a map that holds information of roads and lanes in a mesh for automated driving and driving assistance, a method of constructing a high-precision map by collecting road information by a vehicle for map generation or a person is known. In this method, a considerable amount of human labor is generated, so only information of a part of roads such as an expressway can be constructed, and the coverage of the constructed map information decreases, so there is a problem that the range of providing a map to automated driving and driving assistance narrows.

[0003] In view of this problem, for example, Patent Literature 1 discloses a method of updating a map using data collected from a camera mounted on a general vehicle. Patent Literature 2 discloses a technique of improving the precision of a map using an aerial photograph and navigation map data. Further, Patent Literature 3 discloses a method of correcting a map using aerial photograph data and basic road map data.

[0004] PRIOR ART DOCUMENTS

[0005] PATENT LITERATURE

[0006] Patent Literature 1: Japanese Patent No. 6197393

[0007] Patent Literature 2: Japanese Patent No. 4799709

[0008] Patent Literature 3: Japanese Patent Laid-Open No. 2019-101256 SUMMARY

[0009] PROBLEMS TO BE SOLVED BY THE INVENTION

[0010] However, the conventional technology has a problem that a shielded area in which a road is shielded and does not appear on an image cannot be properly processed.

[0011] In an aerial photograph, a satellite image, or an image given by a camera mounted on a general vehicle, there is a shielded area in which a road is shielded by a tunnel, an overpass, a group of high-rise buildings, or the like. Therefore, when road information extracted from an image captured is reflected on a map, there is a problem that incorrect information is reflected.

[0012] The present application has been made to solve such a problem, and an object thereof is to provide an information processing apparatus and an information processing method that can properly process information of a shielded area of a road.

[0013] Technical means for solving the problem

[0014] An example of the information processing apparatus of the present application includes:

[0015] a first information conversion section that infers a first road region from an image obtained by photographing a road;

[0016] a second information conversion section that infers a second road region from map information stored in a map information storage section; and

[0017] a difference extraction section that extracts difference information indicating a difference between the first road region and the second road region.

[0018] An example of the information processing method of the present application includes,

[0019] inferring a first road region from an image obtained by photographing a road by a computer,

[0020] inferring a second road region from map information,

[0021] and extracting difference information indicating a difference between the first road region and the second road region.

[0022] This specification contains the disclosure of Japanese Patent Application No. 2021-063499, which is a basis for priority of this application.

[0023] Effects of the Invention

[0024] The information processing apparatus and the information processing method of the present application can appropriately process a shielded region of a road. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 A diagram for explaining a basic configuration of the information processing apparatus.

[0026] Figure 2 A diagram for explaining the information conversion section.

[0027] Figure 3 A diagram for explaining the information conversion section having the shielded region inference section.

[0028] Figure 4 A diagram for explaining the difference extraction section.

[0029] Figure 5 A diagram for explaining the map information verification section.

[0030] Figure 6 A diagram for showing a relationship between the difference information and the reliability under the configuration of Figure 5

[0031] Figure 7 ​FIG. 1 is a diagram for explaining a configuration of an information processing apparatus having a map information updating section.

[0032] Figure 8 FIG. 2 is a diagram for explaining a mutual occlusion information verification section.

[0033] Figure 9 FIG. 3 is a diagram for explaining a relationship between differential information and reliability in a configuration of Figure 8

[0034] Figure 10 FIG. 4 is a diagram for explaining a configuration having a differential reliability output section.

[0035] Figure 11 FIG. 5 is a diagram for explaining an example in which different things are output in map information and image information.

[0036] Figure 12 FIG. 6 is a diagram for explaining an example of a method of determining reliability of map information. DETAILED DESCRIPTION

[0037] Hereinafter, an embodiment of the present application will be described with reference to the drawings.

[0038] [Example 1]

[0039] Figure 1 A configuration of the information processing apparatus of Example 1 is shown. The information processing apparatus executes the information processing method described in this specification.

[0040] The image acquisition section 12 acquires a satellite image or an aerial image belonging to a range specified by the latitude / longitude specifying section 13 from the image information storage section 10 that stores satellite images and aerial photographs. Here, the satellite image is, for example, data obtained by acquiring an image by photographing a ground surface with a camera mounted on a satellite and restoring the ground surface in a bird's-eye view. Further, the aerial image is, for example, data obtained by acquiring an image by photographing a ground surface with a camera mounted on an aircraft and restoring the ground surface in a bird's-eye view.

[0041] The latitude / longitude specifying section 13 can specify latitude / longitude in such a manner that an image of an object to be acquired is extracted, or can use an identifier or the like instead of latitude / longitude.

[0042] In the first information conversion section 15, a road region (first road region) is inferred from an image obtained by photographing a road. For example, a satellite image or an aerial image acquired by the image acquisition section 12 is processed to infer a road region. The first road region can be expressed as an image.

[0043] ​The road region is inferred, for example, as an attribute given to each pixel of the image, and contains information indicating whether each pixel is a road. This information can be represented as an image. Alternatively, the extracted road region can be represented in a mesh form, in which case the extracted information can be represented by a set of pixels.

[0044] In the map information acquisition section 14, the map information (for example, information in a mesh form) of the region specified by the latitude / longitude specifying section 13 is acquired from the map information storage section 11.

[0045] The second information conversion section 17 infers a road region (second road region) from the map information stored in the map information storage section 11. For example, data in the same form as the data output by the first information conversion section 15 is output. That is, in the case where the first information conversion section 15 outputs an image, the second information conversion section 17 also generates an image indicating a road region. In the case where the first information conversion section 15 outputs data in a mesh form, the second information conversion section 17 also outputs data in a mesh form.

[0046] The difference extraction section 16 extracts difference information indicating the difference between the above two road regions (first road region and second road region).

[0047] The structure information verification section 18 verifies a structure related to a road from the extracted difference information and outputs the result. Further, in the following, the form output by the first information conversion section 15 and the second information conversion section 17 will be assumed to be an image unless specifically stated otherwise, but the output of both can also be a mesh. The difference extraction section 16 can compare images with each other, or can compare meshes with each other.

[0048] Figure 1 The information processing apparatus can be configured using a known computer, for example. The information processing apparatus has, for example, an arithmetic unit and a storage unit. The arithmetic unit includes, for example, a processor, and the storage unit includes, for example, a semiconductor storage device, a magnetic disk device, and the like.

[0049] Further, the computer can have an input / output unit. The input / output unit includes, for example, an input device such as a keyboard and a mouse, an output device such as a display and a printer, and a communication device such as a network interface.

[0050] The storage unit can store a program. The computer can execute the functions described in the present embodiment by executing the program by the processor. That is, the computer functions as the information processing apparatus by functioning as each of the constituent elements shown in the block diagram, and thereby functions as the information processing method. Figure 1 The computer functions as the information processing apparatus by functioning as each of the constituent elements shown in the block diagram, and thereby functions as the information processing method.

[0051] The configurations of the first information conversion section 15 and the second information conversion section 17 are shown in the block diagram of FIG. 2. Figure 2In the first information conversion section 15, it is determined, for example, whether or not each pixel of the image obtained by the image acquisition section 12 is a road region. Here, the determination can be made by a technique such as semantic segmentation using a previously learned model 150 such as a neural network learned in advance by machine learning.

[0052] Further, in the following description, the term "road" is used as a general term for a road composed of a single lane and a road composed of a plurality of lanes. For example, for a road composed of a plurality of lanes, the entire road can be treated as one road, or the plurality of lanes can be treated as a plurality of roads in parallel.

[0053] The determination of the road region is performed in the first road region determination section 151. In addition, there can be a first road boundary determination section 152 that determines the boundary between a road and another road or between a road and a non-road, and there can be a first road indication region determination section 153 that determines a white line, a yellow line, a zebra crossing, or the like. In order to determine more detailed information, there can be a first lane boundary determination section 154 and / or a first lane region determination section 155 that divides the road region by each lane and determines the boundary of the lane together.

[0054] The map information storage section 11 stores, for example, road data included in map data for navigation. The data here stores the shape of the road in a mesh, for example, with nodes and connecting lines, and states indicating the shape of the connecting lines and the attributes of the road. For example, the end of the road is represented as a node, and the road is represented as a connecting line connecting two nodes. Information indicating the attributes of the road (for example, the width of the road or the lane) can be associated with the connecting line.

[0055] Like the first information conversion section 15, the second information conversion section 17 can have, in addition to the second road region determination section 171, a second road boundary determination section 172 that determines the boundary between a road and a non-road or the boundary between a road and another road, and in order to determine more detailed information, there can be a second lane boundary determination section 174 and / or a second lane region determination section 175 that divides the road region by each lane and determines the boundary of the lane together. In more detail, in the case where the map holds information of a white line, a yellow line, or the like, there can be a second road indication region determination section 173.

[0056] For example, in a case where the mesh information given indicates the position of a center line of a road, the second road region estimation section 171 can first convert the mesh information into image-form information, and further widen the pixels on both sides of the center line to estimate the road region. The widening is performed, for example, by forming a circular plate with a radius of half the width of a lane in consideration of the number of lanes for each pixel constituting the center line, and superimposing the circular plates, thereby widening. The number of pixels to be widened can be determined, for example, from the distance of each pixel of the image obtained by the first information conversion section 15. When the widening is performed, in a case where two parallel roads are repeated, the center line of the repeated region can be regarded as a road boundary. For a case where there is a vertical relationship of heights of roads (for example, an overpass), the processing can be designed as appropriate by those skilled in the art, for example, so that the road with a higher height is given priority in the widening.

[0057] The function of the difference extraction section 16 is shown in Figure 4 . Further, in the present embodiment, the results of the first road region estimation section 151 and the second road region estimation section 171 shown in Figure 2 are used, but the processing results of the first road boundary estimation section 152 and the second road boundary estimation section 172, the first road indication region estimation section 153 and the second road indication region estimation section 173, the first lane boundary estimation section 154 and the second lane boundary estimation section 174, and the like can also be used as information for improving the accuracy of the difference extraction.

[0058] In the difference extraction section 16, the difference between the image obtained from the first information conversion section 15 and the image obtained from the second information conversion section 17 is extracted. The images obtained from the two information conversion sections can have a positional shift of roads from each other due to the distortion of the satellite image or aerial photograph that is the basis thereof, or the shift of the position of the mesh of the map registered in the map information storage section 11. Therefore, first, the rough alignment of roads from each other is performed in the initial position correction section 161. The specific processing of the alignment between the two images can be designed as appropriate according to known techniques.

[0059] To perform the alignment, the second information conversion section 17 can regenerate the image by adjusting with reference to the original mesh information. Next, according to the comparison form (162), in a case where the comparison is performed according to the road, the comparison of the presence / absence of a road on the image, the comparison of the road widths from each other are performed in the road presence / absence determination section 163 and the road width comparison section 165, and in a case where the comparison is performed according to the lane, the comparison of the presence / absence of a lane, the comparison of the lane widths are performed in the lane presence / absence determination section 164 and the lane width comparison section 166.

[0060] In the comparison of the presence or absence of the lane and the comparison of the lane width, a process of identifying the road according to the first road region and the second road region can be performed. Further, after the road is identified, a process of identifying the end point and the width of the road can be performed. Furthermore, a process of determining the correspondence relation of the road between the respective road regions can be performed. These specific processes can be designed as appropriate according to known techniques. Thus, the presence or absence of the road or the width of the road can be compared.

[0061] The road presence / absence determination section 163 determines the difference in the presence or absence of the road within the first road region and the second road region (both of which are images in the following examples). For example, in a case where there is information of a certain road within one image and there is no information of the corresponding road within the other image, it is determined that there is a difference in the presence or absence of the road.

[0062] In a case where there is information of a road associated with both images, the road width comparison section 165 compares the width of the road on the images, calculates the difference in the width, and outputs it. The output of the difference can be processed on the basis of taking into account the conversion error and the like in the input of both data, and the output unit can be considered to be the number of pixels or the number of lanes equivalent thereto or the like.

[0063] The lane presence / absence determination section 164 determines the difference in the presence or absence of the lane within the first road region and the second road region. In a case where there is information of a lane associated with both images, the lane width comparison section 166 compares the width of the lane on the images, calculates the difference in the width, and outputs it.

[0064] The difference extraction section 16 sends these pieces of information to the structure information verification section 18 as difference information. As described above, the difference information includes difference presence / absence information indicating the presence or absence of the difference and difference amount information indicating the amount of the difference. Here, the "presence or absence of the difference" means, for example, whether there is a difference in the presence or absence of the road and whether there is a difference in the width of the road, and the "amount of the difference" means, for example, the magnitude of the difference in the width of the road.

[0065] As the unit of the output difference, for example, a part of the road or the lane, a set of pixels on the image, a node and a connection line in a mesh, or the like can be considered. Hereinafter, the unit of the output difference will be referred to as a division. The division is, for example, a division of an image into a predetermined reference (for example, into a square of an equal size). Further, as to which part of the information in the original image information storage section 10 and the map information storage section 11 the output division corresponds to, it can be determined by tracing the conversion history.

[0066] The configuration of the structure information verification section 18 is shown in Figure 5 In this example, the structure information verification section 18 is provided with a map information verification section 20.

[0067] The map information verification section 20 classifies the kind of difference based on the comparison result of the difference extraction section 16 and the structure information of the road or the lane and outputs it. Further, the function will be explained with respect to the road below, but it can be replaced with the lane. In the structure information verification section 18, it is confirmed in the map information verification section 20 whether or not the shielding occurs in the division which becomes the object.

[0068] In the map information verification section 20, it is determined whether or not there is a shielding area (i.e., whether or not the road is shielded by other structure) with reference to the information of the corresponding map information storage section 11 in the division in which the difference occurs. For example, as exemplified in the above Figure 11 If the road is in the tunnel, it is determined that the part is a shielding area. Further, for example, if the road is an underpass road corresponding to an overpass or an expressway, it is determined that the part is a shielding area. In the part in which the roads are intersected three-dimensionally, the width of both roads can be considered to determine the shielding area.

[0069] Here, as the output, 1 is output if there is shielding and 0 is output if there is no shielding, and it is output in the form of a variable "#" of Figure 5 The value of the variable # can also be set to levels in stages according to the degree of shielding or the like.

[0070] First, in the case where it is determined that there is a difference in the difference extraction section 16 (S1: Yes), if there is a road on the image given from the first information conversion section 15 and there is no road on the image given from the second information conversion section 17 (S2), a signal A is output as the classification result.

[0071] In the case where it is determined that there is a difference (S1: Yes), there is no road on the image given from the first information conversion section 15, and there is a road on the image given from the second information conversion section 17 (S3), a signal B is output. In particular, in the map information verification section 20, for the road involved in the difference, a signal B0 is output if there is no shielding and a signal B1 is output if there is shielding.

[0072] In the case where it is determined that there is a difference (S1: Yes), there is a road on the image given from the first information conversion section 15 and there is a road on the image given from the second information conversion section 17 (S4), a signal C or D is output. That is, the case where there are roads on both images but the number of lanes or the road width or the like registered on the road is different.

[0073] In this case, if the width (or the number of lanes) of the road observed from the image given from the first information conversion section 15 is larger than that given from the second information conversion section 17 (S5: Yes), a signal D is output in the map information verification section 20, and otherwise (S5: No), a signal C is output.

[0074] Specifically, for signal C, in the map information verification unit 20, if the road involving the difference is unobstructed, signal C0 is output; if it is obstructed, signal C1 is output. Similarly, for signal D, in the map information verification unit 20, if the road involving the difference is unobstructed, signal D0 is output; if it is obstructed, signal D1 is output.

[0075] If the difference extraction unit 16 determines that there is no difference (S1: No), then when there is a road in both images (S6), signal E is output. In particular, signal E0 is output when there is no occlusion, and signal E1 is output when there is occlusion.

[0076] In the case where the differential extraction unit 16 determines that there is no difference (S1: No) and there is no road on the two images (S7), the output signal F is generated.

[0077] Thus, the structural information verification unit 18 distinguishes the types of differences and outputs the results by confirming the occlusion status on the map where the differences have occurred.

[0078] Information processing devices can be like Figure 10 As shown, it has a differential reliability output unit 30, which can be configured to output a differential reliability output unit 30 according to... Figure 5 The reliability of the difference is determined by the results of the processing.

[0079] Figure 6 The relationship between differential information and reliability is demonstrated. The differential reliability output unit 30 outputs reliability for both the presence / absence of differential information and the differential component information. In this embodiment, the differential reliability output unit 30 determines the reliability based on the first road area, the second road area, and the presence or absence of obstruction. With this configuration, the presence or absence of obstruction can be appropriately considered when determining the reliability.

[0080] The reason why the reliability is set to "medium" in case A is that there is no data for reference on the map information storage unit 11 side.

[0081] The "-" indicates a situation where there is no data to update due to the lack of difference. The reliability under occlusion (B1, C1, D1, E1) is lower than that under no-occlusion (B0, C0, D0, E0). This is because occlusion can potentially lead to inaccurate road identification using images.

[0082] For road recognition using images (based on processing of a first road region), if occlusion exists, a portion of the road will not be represented in the image. Therefore, the identified road is more likely to be narrower than the actual road and less likely to be wider. Consequently, the reliability is low in case C1 and high in case D1 (higher than C1).

[0083] Information processing devices can be like Figure 7As shown, the map information updating section 31 is provided. The map information updating section 31 updates the map information of the map information storage section 11 according to the differential information (for example, according to the results of the structure information verification section 18 and the differential reliability output section 30). More specifically, whether to update the map information is decided according to the reliability of the differential presence / absence information and the reliability of the differential amount information. Also, in the case where it is decided to update, the map information is updated. According to such a configuration, the map information can be properly updated.

[0084] In the present embodiment, the update of the map information is performed according to the results of the classification in Figure 5 and Figure 6

[0085] Further, the following shows a reference example of the update, assuming that the data of the image information storage section 10 is new compared to the data of the map information storage section 11. In the case where the data of the image information storage section 10 is old, the map information updating section 31 does not perform the update of the map information. However, as a modification example, the update can also be performed in the case where the data of the image information storage section 10 is old. Further, as another modification example, the map information can also be updated in the case where the newness / oldness of the data is similar (for example, the time difference is within a prescribed range) or the reliability of the map information is low. Note that the "reliability of the map information" can be determined by the processing described later, or can be determined using other appropriate reference.

[0086] First, an example of additionally registering whether the reliability is high or low when registering the information in the map will be shown. Figure 5 , Figure 6 In the case of A (that is, the case where a road exists within the first road region and a corresponding road does not exist within the second road region), the map information updating section 31 can decide to update the map information. As a specific example, the road-related information can be added to the map information. In this way, in the case where a new road is found by means of the image, the information can be reflected in the map information. Note that as a modification example, the map information can also be decided not to be updated in the case of A.

[0087] In the case of B0, C0, D0, and D1, which are considered to have a high possibility of being correct, the map is given information that the reliability of the map information is low. On the other hand, in the case of B1, C1, and E1, which are considered to have a high possibility of being incorrect, the map is given information that the map information is not updated or that the reliability of the map information is high as a result of the verification by the image information storage section 10.

[0088] ​That is, in a case where there is a road in the second road region and there is no corresponding road in the first road region (B), the map information updating section 31 can decide not to update the map information when it is determined that the difference information is generated due to the occlusion of the road (Bl). On the other hand, the map information updating section 31 can decide to update the map information when it is not determined that the difference information is generated due to the occlusion of the road (B0). In this way, it is possible to appropriately update the map information taking into account the presence or absence of the occlusion.

[0089] Further, in a case where there is a corresponding road in both the first road region and the second road region (Cl and Dl), the map information updating section 31 further decides whether to update the map information (Cl or Dl) depending on the width of the road in the first road region and the width of the road in the second road region. In this way, it is possible to appropriately update the map information taking into account the width of the road.

[0090] For E0 and F, information is given that there is no need to update the reliability, or in the case of E0, information is given that the result of the verification of the image information storage section 10 is that the reliability of the map information is high. For A, it is possible to decide to update the map information, or it is possible to give information that there is no need to update the map information because there is no road information on the map that should be updated. Or, for A, for example, in a case where the road where the difference has occurred intersects with another road, it is possible to give information that the reliability is low. At this time, it is also possible to newly add an update node and a connecting line for the intersection to the mesh information.

[0091] Next, a case where the reliability of the number of lanes and the width of the lanes registered in the map information is high or low is considered. Here, in a case where the result of the reliability of the difference amount is high in Figure 6 , it is considered that the reliability of the information registered in the map is low, and it is considered to give the result. In Figure 6 , it is considered that the information registered in the map has reliability, and it is considered to give information that there is no need to update the map information or to give information that the reliability is high.

[0092] In the above, a method of determining the reliability of the map information based on the reliability of the difference is described, but of course it is also possible to directly register the information of A to F as information indicating the reliability of the map information.

[0093] As described above, according to the information processing apparatus of the present embodiment, it is possible to appropriately handle the occluded region of the road. For example, by adopting a method based on the presence or absence of the difference and the reliability of the difference with respect to the output, it is possible to update the map information taking into account the occlusion information which is a problem unique to satellite images and aerial photographs.

[0094] Figure 3 The display in Figure 2The first information conversion section 15 and the second information conversion section 17 are additionally provided with a function section that estimates a shielded area. The first information conversion section 15 is provided with a first shielded area estimation section 156, and the second information conversion section 17 is provided with a second shielded area estimation section 176.

[0095] The first information conversion section 15 estimates first structure information from the image with the first shielded area estimation section 156. The first structure information is, for example, information that specifies a portion of the image in which a road is shielded (shielded area). The specific processing of the estimation can be designed as appropriate according to known techniques and the like. At the time of the estimation, the previously learned model 150 can estimate a shielded road area from the surrounding building group or pattern information.

[0096] There are cases in which a shielded area is shielded by a structure other than a road, such as a building or a tree, and cases in which a shielded area is shielded by another road, such as an overpass or an expressway. Therefore, a plurality of attributes can be output for one pixel. For example, the attribute of a structure other than a road and the attribute of a shielded road can be output, and the attribute of a road and the attribute of a shielded road can be output.

[0097] In addition, an output that estimates the cause of the shielding, such as shielding by a building, shielding by a tree, and shielding by another road, can be provided.

[0098] The second information conversion section 17 estimates second structure information from the map information with the second shielded area estimation section 176. The second structure information is, for example, information that specifies a portion of the map information in which a road is shielded. This estimation can be performed by the same processing as the map information verification section 20. In the present embodiment, the difference extraction section 16 extracts the difference in the road area, not the difference in the shielded area.

[0099] There are several differences between the shielded areas estimated by the first shielded area estimation section 156 and the second shielded area estimation section 176. One difference is that the shielded areas that the shielded area estimation section of the first information conversion section 15 is able to estimate tend to be shorter in distance or narrower in area compared to the shielded areas estimated by the second information conversion section 17. For example, the first information conversion section 15 is able to estimate shielding for a road in which a very small portion of the road is shielded, but is not able to appropriately estimate shielding for an area in which a large portion of the road is shielded, such as a long tunnel.

[0100] In such a configuration in which the first shielded area estimation section 156 and the second shielded area estimation section 176 are provided, the configuration of the structure information verification section 18 is, for example, provided with a mutual shield information verification section 21. Figure 8 In the mutual shield information verification section 21, the state of the shielding is estimated by comparing the first structure information and the second structure information.

[0101] In the mutual occlusion information verification section 21, if there is an occlusion in both the first structure information (image) and the second structure information (map information), 3 is output as the variable "$" of Figure 8 , if there is an occlusion only in the map information, 2 is output as the variable "$" of Figure 8 , if there is an occlusion only in the image, 1 is output as the variable "$" of Figure 8 , and if there is no occlusion in either, 0 is output as the variable "$" of Figure 8 . The meaning of the variable "$" is the same as that of the variable # of Figure 5 , but the variable "$" can take 4 values from 0 to 3. According to such a configuration, the occlusions of the image and the map information can be considered.

[0102] In view of such a feature, the structure information verification section 18 verifies the reliability of the extracted difference region. As a specific example, in the case where a difference is extracted in the difference extraction section 16, the possibility of occlusion is judged, and in the case where the entire road of one side is occluded, the map information is not updated. Further, in the case where only a part is occluded, the reliability of the difference is lowered. On the other hand, in the case where no occlusion is judged, the reliability of the difference is increased.

[0103] The classification of A to F shown above is to classify the reliability of the extracted difference and whether the amount of the difference can be trusted, in addition to the presence or absence of the road indicated by the extracted difference, on the premise that there is an occlusion in the image information storage section 10.

[0104] As in Figure 6 , for the configuration of Figure 8 , the reliability can be determined as in Figure 9 . For A, since there is no occlusion information corresponding to the map side, there are no A3 and A2. Similarly, F has no F3 and F2. The result of such classification can be output as the output of the difference reliability output section 30 shown in Figure 10 .

[0105] The map information can also be updated according to the reliability shown in Figure 9 . The specific processing at this time can be designed in the same manner as the example explained in association with Figure 6 .

[0106] Using Figure 12 , an example of a case where only the information of the image information storage section 10 is updated, and the map information is not updated, after the update under the configuration of Figure 7 , is described. Here, the reliability of the map information takes values from 0 to 3, and the larger the value, the higher the reliability of the map information. Further, this "reliability of the map information" is different information from the above-mentioned "reliability of the difference".

[0107] Further, the case where the data (image) of the image information storage section 10 is newer than the data (map information) of the map information storage section 11 is considered here, but the following processing can be applied equally even in a case where this is not true.

[0108] In a case where it is determined that there is no difference (Sll: No), the map information updating section 31 sets the reliability of the map information to 3. In a case where it is determined that there is a difference (Sll: Yes), the following processing is performed.

[0109] In a case where it is determined that the reliability of the difference is low (S12: No), the map information updating section 31 sets the reliability of the map information to 2. In a case where it is determined that the reliability of the difference is high (S12: Yes), the following processing is performed.

[0110] The map information updating section 31 refers to the reliability of the last time (that is, the reliability of the map information determined when the map information was last updated). In a case where the reliability of the last time is not specified (S13: No), the reliability of the map information is set to 1. In a case where the reliability of the last time is 0 or 1 (S14: Yes), the reliability of the map information is set to 0, and in a case where the reliability of the last time is 2 or 3 (S14: No), the reliability of the map information is set to 1.

[0111] According to this example, the map information updating section 31 determines the reliability of the map information when the map information is updated. Then, the new reliability of the map information is determined based on the reliability of the map information determined when the map information was last updated. Therefore, the reliability of the map information can be determined more appropriately based on the reliability of the difference, the reliability of the map information, and the history of the reliability of the map information. The reliability of the map information is used as a criterion for determining whether or not to update the map information, for example, so the map information can be updated more appropriately.

[0112] Next, an example of a case where the data (map information) of the map information storage section 11 is newer than the data (image) of the image information storage section 10 will be described. Sometimes, the update of the map information is performed only for the nodes and connection lines for which a difference from the last map information has occurred. In this case, the last difference information relating to the nodes and connection lines to be updated is discarded and only the result of the new comparison is registered.

[0113] Finally, in a case where the newness of the data of both is the same (for example, in a case where the update is performed at the same time), the last difference information is discarded for the part of the nodes and connection lines for which the map information has been updated and the new comparison result is registered, and the update rule shown in Figure 12 is applied for the other part.

[0114] Symbol Explanation

[0115] 15 first information conversion section, 16 difference extraction section, 17 second information conversion section, 18 structure information verification section, 30 difference reliability output section, 31 map information update section

[0116] All publications, patents, and patent applications cited in this specification are herein incorporated by reference in their entirety.

Claims

1. An information processing apparatus, characterized by, Possessing: a first information conversion section that infers a first road region from an image obtained by photographing a road; a second information conversion section that infers a second road region from map information stored in a map information storage section; a difference extraction section that extracts difference information indicating a difference between the first road region and the second road region; a difference reliability output section that determines a reliability of the difference information; and a map information update section that decides whether or not to update the map information based on the difference information, in a case where there is a road in the second road region and there is no corresponding road in the first road region, in a case where it is determined that the difference information is generated due to a road occlusion, the map information update section decides not to update the map information, in a case where it is not determined that the difference information is generated due to a road occlusion, the map information update section decides to update the map information.

2. The information processing apparatus according to claim 1, wherein the difference information includes difference presence / absence information indicating a presence / absence of a difference and difference amount information indicating an amount of a difference, the difference reliability output section outputs a reliability for the difference presence / absence information and the difference amount information, respectively, the map information update section decides whether or not to update the map information based on the reliability of the difference presence / absence information and the reliability of the difference amount information.

3. The information processing apparatus according to claim 1, wherein the map information update section determines a reliability of the map information when updating the map information.

4. The information processing apparatus according to claim 3, wherein the map information update section determines the reliability of the map information based on a reliability of the map information determined when the map information was last updated.

5. The information processing apparatus according to claim 1, wherein the first information conversion section infers first structure information from the image, the second information conversion section infers second structure information from the map information, the information processing apparatus possesses a structure information verification section that collates the first structure information and the second structure information, the first structure information is information that determines a portion of an image in which there is a road occlusion, and the second structure information is information that determines a portion of map information in which there is a road occlusion.

6. The information processing apparatus according to claim 1, wherein in a case where there is a road in the first road region and there is no corresponding road in the second road region, the map information update section decides to update the map information.

7. The information processing apparatus according to claim 1, wherein in a case where there is a corresponding road in both the first road region and the second road region, the map information update section further decides whether or not to update the map information based on a width of a road in the first road region and a width of a road in the second road region.

8. An information processing method, characterized by inferring, by a computer, a first road region from an image obtained by photographing a road, ​ ​ inferring a second road region from map information, and extracting difference information representing a difference between the first road region and the second road region, in a case where there is a road in the second road region and there is no corresponding road in the first road region, in a case where it is determined that the difference information is generated due to occlusion of a road, deciding not to update the map information, in a case where it is not determined that the difference information is generated due to occlusion of a road, deciding to update the map information.

Images