JP6318962B2 - Image generating apparatus and image generating method - Google Patents

Description

  The present invention relates to an image generation apparatus and method for generating an image from a radio wave image generated by detecting electromagnetic waves coming from the front of a vehicle and a visible image obtained by imaging the front of the vehicle.

  Conventionally, Patent Document 1 is disclosed as a technique for generating a composite image by combining a plurality of images.

  In the imaging apparatus disclosed in Patent Document 1, when the sensitivity of the imaging element is low, which is lower than a threshold value, a plurality of images that are captured a plurality of times with an exposure amount smaller than the standard exposure are combined to obtain a single sheet. An image for noise reduction was generated.

JP 2012-239077 A

  However, in the above-described conventional imaging apparatus, the contrast on the image is lowered even if the images are combined at night or in backlight. Therefore, when mounted on a vehicle, there has been a problem that an image obtained by imaging a road surface necessary for traveling of the vehicle cannot be provided to the driver.

  Therefore, the present invention has been proposed in view of the above-described circumstances, and can provide a driver with an image obtained by capturing the road surface necessary for driving the vehicle even at night or in backlighting. An object of the present invention is to provide a generation apparatus and a method thereof.

  In order to solve the above-described problem, an image generation apparatus and method according to an aspect of the present invention generate a radio wave image by receiving horizontal polarization components and vertical polarization components of electromagnetic waves arriving from the front of a vehicle, An object region is set by classifying objects on a radio wave image based on horizontal polarization components and vertical polarization components. In addition, a low-contrast region where the contrast is lowered on the visible image is detected by capturing a visible image in front of the vehicle. And the output image which improved the contrast of the area | region corresponding to the object area | region on a radio wave image among the low contrast area | regions on a visible image is produced | generated.

  ADVANTAGE OF THE INVENTION According to this invention, even when the contrast on an image falls at night, a backlight, etc., the image which imaged the state of the road surface required for driving | running | working of a vehicle can be provided to a driver | operator.

FIG. 1 is a block diagram showing a configuration of an image generation apparatus according to the first embodiment of the present invention. FIG. 2 is a flowchart showing a processing procedure of image generation processing by the image generation apparatus according to the first embodiment of the present invention. FIG. 3 is a diagram illustrating a radio wave image and a visible image generated and captured by the image generation apparatus according to the first embodiment of the present invention. FIG. 4 is a diagram for explaining an object classification method by the image generation apparatus according to the first embodiment of the present invention. FIG. 5 is a diagram illustrating an output image generated by the image generation apparatus according to the first embodiment of the present invention. FIG. 6 is a flowchart showing a processing procedure of image generation processing by the image generation apparatus according to the second embodiment of the present invention. FIG. 7 is a diagram illustrating a radio wave image and a visible image generated and captured by the image generation apparatus according to the second embodiment of the present invention. FIG. 8 is a diagram illustrating a radio wave image and a visible image generated and captured by the image generation apparatus according to the second embodiment of the present invention. FIG. 9 is a diagram illustrating a relationship between a pixel position and a luminance value of a visible image captured by the image generation device according to the second embodiment of the present invention. FIG. 10 is a diagram illustrating an output image generated by the image generation apparatus according to the second embodiment of the present invention. FIG. 11 is a diagram illustrating a relationship between a pixel position and a luminance value of a visible image captured by the image generation device according to the second embodiment of the present invention. FIG. 12 is a diagram illustrating an output image generated by the image generation apparatus according to the second embodiment of the present invention. FIG. 13 is a block diagram showing a configuration of an image generation apparatus according to the third embodiment of the present invention. FIG. 14 is a flowchart showing a processing procedure of lane estimation processing by the image generation apparatus according to the third embodiment of the present invention. FIG. 15 is a diagram illustrating a radio wave image and a visible image generated and captured by the image generation apparatus according to the third embodiment of the present invention. FIG. 16 is a diagram illustrating an output image generated by the image generation apparatus according to the third embodiment of the present invention.

  Hereinafter, first to third embodiments to which the present invention is applied will be described with reference to the drawings.

[First Embodiment]
[Configuration of Image Generation Device]
FIG. 1 is a block diagram illustrating a configuration of an image generation apparatus according to the present embodiment. As shown in FIG. 1, the image generation apparatus 1 according to the present embodiment includes a radio wave reception unit 3, a visible image imaging unit 5, an object classification unit 7, a contrast detection unit 9, and an image generation unit 11. ing.

  Here, the image generation device 1 according to the present embodiment is mounted on a vehicle, generates an image in front of the vehicle, and provides the image to the driver. At this time, when the contrast of the image is reduced due to nighttime or backlight, the output image is generated by improving the contrast of the area where the road surface or the three-dimensional object is imaged among the areas where the contrast is reduced.

  The radio wave receiving unit 3 includes a horizontal polarization antenna that mainly detects a horizontal polarization component of electromagnetic waves arriving from the front of the vehicle, and a vertical polarization antenna that detects mainly a vertical polarization component of electromagnetic waves arriving from the front of the vehicle. A radio wave image is generated from the horizontal polarization component and the vertical polarization component received by these antennas. Here, when generating the radio wave image, the radio wave receiving unit 3 can generate a reference image generated at a normal resolution and a high resolution image generated at a high resolution. For example, in the normal state, the reference image is generated by scanning the antenna while skipping one pixel at a time, and when necessary, the high-resolution image is generated by scanning all the pixels of the antenna. In particular, in the present embodiment, a high-resolution radio wave image is generated for an area classified as a road surface area by the object classification unit 7.

  The visible image capturing unit 5 is a normal visible camera for capturing a visible image in front of the vehicle.

  The object classification unit 7 classifies the object on the radio wave image generated by the radio wave reception unit 3 based on the horizontal polarization component and the vertical polarization component. More specifically, the object classification unit 7 calculates a polarization ratio indicating the ratio between the horizontal polarization component and the vertical polarization component in each pixel on the radio wave image generated by the radio wave reception unit 3. The object captured in the radio wave image is classified based on the polarization ratio and the intensity of the vertical polarization component. For example, the object classification unit 7 includes a table that is set in advance according to the polarization ratio and the intensity of the vertical polarization component. Using this table, each pixel on the radio wave image is displayed in the sky, a three-dimensional object, or a road surface. Classify as rough terrain. Then, each classified area is set as each object area. For example, an area classified as a road surface is set as a road surface area.

  The contrast detection unit 9 detects a low contrast region in which the contrast is reduced on the visible image captured by the visible image capturing unit 5. The contrast detection unit 9 detects the luminance of each pixel of the visible image and calculates the standard deviation of the luminance in each row of the visible image. Then, an area where the standard deviation is a predetermined value or less is detected as a low contrast area.

  The image generation unit 11 generates an output image by improving the contrast of the region corresponding to the road surface region on the radio wave image among the low contrast regions on the visible image. In the present embodiment, an output in which the contrast of the road surface area is improved by synthesizing the high-resolution radio wave image generated by the radio wave receiver 3 in an area corresponding to the road surface area in the low contrast area on the visible image. Generate an image.

  Here, the image generation apparatus 1 includes a general-purpose electronic circuit including a microcomputer, a microprocessor, and a CPU, and peripheral devices. Then, by executing a specific program, the radio receiver 3, the visible image capturing unit 5, the object classification unit 7, the contrast detection unit 9, and the image generation unit 11 are operated.

[Image generation processing procedure]
Next, the procedure of image generation processing by the image generation apparatus 1 according to the present embodiment will be described with reference to the flowchart of FIG.

  As shown in FIG. 2, first, in step S101, the radio wave receiver 3 receives a horizontal polarization component and a vertical polarization component of electromagnetic waves coming from the front of the vehicle and generates a radio wave image. For example, the radio wave receiving unit 3 generates a radio wave image as shown in FIG. In the radio wave image of FIG. 3A, a sky 31, a flat road surface 33 such as asphalt, and an irregular ground 35 such as a green zone are captured.

  In step S102, the visible image capturing unit 5 captures a visible image in front of the vehicle. For example, the visible image capturing unit 5 captures a visible image as illustrated in FIG. In the visible image of FIG. 3B, a bright part illuminated by a headlight is imaged on the front side, and darkness at night is imaged beyond that.

In step S <b> 103, the object classification unit 7 calculates a polarization ratio indicating a ratio between a horizontal polarization component and a vertical polarization component in each pixel of the radio wave image generated by the radio wave reception unit 3. This polarization ratio shows the ratio of the horizontal polarization component intensity to the vertical polarization component intensity,
Polarization ratio = Horizontal polarization component intensity / Vertical polarization component intensity (1)
Represented as:

  When the polarization ratio is calculated, the object classification unit 7 classifies the object captured in the radio wave image based on the calculated polarization ratio and the vertical polarization component intensity. Here, as shown in FIG. 4, the object classification unit 7 is provided with a table set in advance according to the polarization ratio and the vertical polarization component intensity, and each pixel on the radio wave image is determined using this table. Classify into sky, solid objects, road surface, rough terrain, etc.

  First, the object classification unit 7 determines which of the conditions 0 to 5 applies to each pixel on the radio wave image using the table in FIG. That is, when the vertical polarization component intensity of each pixel is less than 80% of the black body radiation, it is determined that the condition is zero. Similarly, if the vertical polarization component intensity is 80% or more and less than 85% of black body radiation, condition 1 is applied. If 85% or more and less than 90% is applied, condition 2 is applied. If 90% or more and less than 95% is applied, the condition is applied. 3. If it is 95% or more and less than 100%, condition 4 is determined, and if it is 100%, condition 5 is determined.

  Next, the object classification unit 7 determines which of the conditions a to e applies to each pixel on the radio wave image using the table of FIG. That is, when the polarization ratio of each pixel is less than 0.8, the condition a is determined. Similarly, if the polarization ratio is less than 0.9 and not less than 0.8, condition b, if not less than 0.9 and not more than 1.1, condition c, if greater than 1.1 and not more than 1.2 If condition d is greater than 1.2, it is determined as condition e.

  Then, the object classification unit 7 obtains a combination of the conditions 0 to 5 and the conditions a to e for each pixel on the radio wave image, and classifies the object using the table of FIG. That is, in the case of condition 1 and condition d, it is determined that the pixel is a three-dimensional object. Similarly, in the case of condition 3 and condition a, it is determined as a road surface, in the case of condition 1 to 5 and condition c, it is determined as rough terrain, and in the case of condition 0 and condition c, it is determined as empty. As a result, the object classification unit 7 classifies each pixel of the radio wave image of FIG. 3A into the sky 31, the road surface 33, and the rough terrain 35, and further includes a row including the road surface 33, that is, a lateral direction including the road surface 33. The area is set as the road surface area 37. Therefore, in FIG. 3A, the area from the S1 line to the Sn line including the road surface 33 is set as the road surface area 37. In this way, the object classification unit 7 classifies the object imaged on the radio wave image.

  In step S104, the contrast detection unit 9 detects a low-contrast region where the contrast is reduced on the visible image. First, the contrast detection unit 9 detects the luminance of each pixel of the visible image, and calculates the standard deviation of luminance in the horizontal direction of the visible image, that is, in each row. Then, a row in which the standard deviation is equal to or smaller than a predetermined value is detected as a low contrast region. For example, in the visible image of FIG. 3B, the region from the Sj line to the Sx line located above the bright portion illuminated by the headlight is detected as the low contrast region 39.

  In step S105, the image generation unit 11 detects an area corresponding to the road surface area on the radio wave image among the low contrast areas on the visible image. For example, in FIG. 3, the region corresponding to the road surface region 37 in the low contrast region 39 is a region 41 from the Sj line to the Sn line.

  In step S <b> 106, the radio wave receiving unit 3 generates a high-resolution radio wave image for the road surface area 37. Normally, the radio wave receiving unit 3 generates a reference image by scanning the antenna while skipping one pixel at a time. However, when generating a high resolution radio wave image, all pixels of the antenna are scanned to generate a high resolution image. Generate. Note that a high resolution image may be generated only for the area 41 instead of generating a high resolution image for the entire road surface area 37.

  In step S107, the image generation unit 11 generates an output image in which the contrast of the region corresponding to the road surface region on the radio wave image in the low contrast region on the visible image is improved with respect to the visible image. In particular, in this embodiment, the high-resolution radio wave image generated in step S106 is synthesized with the region corresponding to the road surface region in the low contrast region on the visible image to improve the contrast of the region corresponding to the road surface region. Yes. For example, the image generation unit 11 combines the high-resolution image of the region 41 in the radio wave image of FIG. 3A with the visible image of FIG. 3B to generate the output image shown in FIG. . As a result, it is possible to complement a region that has been lost due to a decrease in contrast on the visible image with the radio wave image to make it easier to see. When the output image is generated in this way, the image generation process according to the present embodiment is terminated.

[Effect of the first embodiment]
As described above in detail, the image generation apparatus 1 according to the present embodiment detects a low-contrast area where the contrast is reduced on the visible image, and detects the low-contrast area in the road image area of the low-contrast area. An output image in which the contrast of the corresponding region is improved is generated. Thereby, even if the contrast on the image is lowered at night or in backlight, an image obtained by imaging the road surface necessary for traveling of the vehicle can be provided to the driver.

  Further, the image generation apparatus 1 according to the present embodiment generates a high-resolution radio wave image for the road surface area, and synthesizes the high-resolution radio wave image in an area corresponding to the road surface area in the low-contrast area on the visible image. Generate an output image. Thereby, even if the contrast on the image is lowered at night or in backlight, an image obtained by capturing the road surface necessary for traveling of the vehicle by the high-resolution radio wave image can be provided to the driver.

[Second Embodiment]
Next, an image generation apparatus according to a second embodiment of the present invention will be described with reference to the drawings. However, since the configuration of the image generation apparatus according to the present embodiment is the same as that of the first embodiment, detailed description thereof is omitted.

[Image generation processing procedure]
Next, the procedure of image generation processing by the image generation apparatus according to the present embodiment will be described with reference to the flowchart of FIG. However, detailed description of steps in which the same processing as in the first embodiment is performed is omitted.

  In the flowchart shown in FIG. 6, in steps S201 and S202, the same processing as that in steps S101 and S102 in FIG. 2 is executed to generate a radio wave image and capture a visible image.

  In step S203, the object classification unit 7 classifies the object captured in the radio wave image as in step S103 in FIG. 2, and sets the area classified as the road surface as the road surface area. For example, as shown in FIG. 7A, a road surface area 71 is set on the radio wave image. Furthermore, in this embodiment, a three-dimensional object detection region is set in a region classified as a three-dimensional object on a radio wave image. For example, as shown in FIG. 8A, solid object detection areas 81 and 83 are set on the radio wave image.

  In step S204, the contrast detection unit 9 detects the luminance of each pixel on the visible image in the same manner as in step S104 of FIG. 2, and detects a low contrast region in which the contrast is lowered. In this embodiment, a high-contrast region with higher contrast is also detected. For example, as shown in FIG. 7B, a low contrast region 73 and a high contrast region 75 are set on the visible image. Furthermore, in this embodiment, an area corresponding to the three-dimensional object detection area set in step S203 is set on the visible image, and the contrast of this area is detected. For example, as shown in FIG. 8B, areas 85 and 87 corresponding to the three-dimensional object detection areas 81 and 83 in FIG. 8A are set on the visible image, and the contrast of these areas 85 and 87 is detected. .

  Next, in step S205, the same processing as in step S105 of FIG. 2 is executed, and an area corresponding to a road surface area on the radio wave image is detected among low contrast areas on the visible image.

  In step S206, the visible image capturing unit 5 captures the visible image by increasing the exposure amount so that the luminance of the region corresponding to the road surface region in the low contrast region of the visible image is increased. However, in FIG. 7, the low-contrast region 73 of the visible image substantially corresponds to the road surface region 71, so that the visible image may be captured by increasing the exposure amount so that the luminance of the low-contrast region 73 increases. If the luminance of the low contrast region can be increased, the exposure amount may not be increased, but the gain may be increased, or both the exposure amount and the gain may be increased.

  Here, when the luminances of the low contrast region 73 and the high contrast region 75 are compared, a result as shown in FIG. 9 can be obtained. FIG. 9 is a graph showing the relationship between the pixel position in the row direction of the visible image and the luminance value at the pixel position. The standard deviation is 11 in the low contrast region 73 and the standard deviation is 74 in the high contrast region 75. It has become.

  Therefore, the visible image capturing unit 5 captures a visible image by increasing the exposure amount so that the standard deviation of the low contrast region 73 increases to the standard deviation of the high contrast region 75. Further, instead of increasing to the standard deviation of the high contrast region, a visible image may be captured by increasing the exposure amount and gain by a preset ratio, for example, three times. As a result, the captured visible image becomes an image as shown in FIG. In the visible image of FIG. 7B imaged by a normal method, the low contrast region 73 was dark because the exposure amount was determined according to the high contrast region 75 by backlight. However, in FIG. 10, since the exposure amount is determined according to the low contrast region 73, the luminance of the low contrast region 73 is increased and the contrast is improved.

  Further, the visible image capturing unit 5 may capture a visible image by increasing the exposure amount so that the luminance of the region corresponding to the three-dimensional object detection region is increased. Here, the luminance of the region 85 corresponding to the three-dimensional object detection region 81 is shown in FIG. FIG. 11 is a graph showing the relationship between the pixel position in the row direction of the visible image including the area 85 and the luminance value at the pixel position. The range indicated by reference numeral 90 indicates the luminance value of the area 85. Yes. In the case of FIG. 11, the standard deviation of the region 85 is 12.

  Therefore, the visible image capturing unit 5 captures a visible image by increasing the exposure amount so that the standard deviation of the region 85 increases to the standard deviation of the high contrast region. As a result, a visible image as shown in FIG. 12 can be taken. In the visible image of FIG. 8B imaged by a normal method, the exposure amount is determined in accordance with the high contrast region by the backlight, so the region 85 where the three-dimensional object is imaged is dark. However, in FIG. 12, since the exposure amount is determined in accordance with the area 85, the brightness of the area 85 is increased and the contrast is improved.

  Thus, when a visible image is captured in step S206, in step S207, the image generation unit 11 generates a visible image (luminance-enhanced visible image) captured so as to increase the luminance as an output image, and relates to the present embodiment. The image generation process ends.

[Effects of Second Embodiment]
As described above in detail, the image generation apparatus according to the present embodiment captures a visible image so that the luminance of the region corresponding to the road surface region in the low-contrast region on the visible image is increased. Generate as output image. As a result, even if the contrast on the image decreases at night or in backlighting, a visible image (luminance-enhanced visible image) in which the luminance of the region corresponding to the road surface region is increased is necessary for driving the vehicle. An image obtained by imaging the road surface state can be provided to the driver.

[Third Embodiment]
Next, an image generation apparatus according to a third embodiment of the present invention will be described with reference to the drawings. However, the same components as those in the first embodiment described above are denoted by the same reference numerals and detailed description thereof is omitted.

[Configuration of Image Generation Device]
FIG. 13 is a block diagram showing the configuration of the image generation apparatus according to this embodiment. As shown in FIG. 13, the image generation device 21 according to the present embodiment is different from the first and second embodiments in that the image generation unit 11 further includes a lane estimation unit 23.

  The lane estimation unit 23 sets a lane detection area on the visible image captured by the visible image capturing unit 5, and estimates the position of the lane from the image analysis result of the lane detection area. At this time, when at least a part of the lane detection area is set in the low contrast area on the visible image, the image generation unit 11 generates an output image in which the contrast of the area including at least the lane detection area is improved. The lane estimating unit 23 sets a lane detection region on the output image and estimates the lane position. The output image in which the contrast of the region including at least the lane detection region is improved may be an output image generated by synthesizing the radio wave image with the visible image in the image region including at least the lane detection region, The output image may be replaced with a radio wave image in the entire image area.

[Procedure for lane estimation processing]
Next, the procedure of the lane estimation process by the lane estimation unit 23 of the image generation device 21 according to the present embodiment will be described with reference to the flowchart of FIG.

  As shown in FIG. 14, in step S301, the lane estimation unit 23 sets a lane detection region on the visible image. For example, as shown in FIG. 15A, a plurality of lane detection areas 101 on a rectangle are set at positions where it is estimated that there is a road surface.

  In step S302, the lane estimation unit 23 determines whether or not at least a part of the lane detection area is set in the low contrast area. Since the low contrast region is detected by the contrast detection unit 9, the lane estimation unit 23 determines whether or not a lane detection region is set in the region. And as shown to Fig.15 (a), when the lane detection area | region 101 is at least partially set to the low contrast area | region 103, it progresses to step S303, and when not setting, it progresses to step S304.

  In step S303, the image generation unit 11 generates an output image by synthesizing the radio wave image with the visible image in an image area including at least the lane detection area, and the lane estimation unit 23 generates an output image on which the radio wave image is combined. Set the lane detection area to. For example, when the radio wave image is replaced in the entire image area, the lane detection area 105 is set in the area on the radio wave image corresponding to the low contrast area 103 as shown in FIG.

  In step S304, the lane estimation unit 23 estimates the position of the lane and displays it on the output image of the image generation unit 11. If it is determined in step S302 that no lane detection area is set in the low contrast area, image analysis is performed on the lane detection area set on the visible image to estimate the position of the lane.

  On the other hand, if it is determined in step S302 that the lane detection area is set in the low contrast area, image analysis is performed on the lane detection areas set on the visible image and on the output image (on the radio wave image). Go and estimate the position of the lane. For example, image analysis is performed on the lane detection area 101 set on the visible image in FIG. 15A, and an image is also displayed on the lane detection area 105 set on the radio wave image in FIG. Analyze and estimate the lane position. As a result, as shown in FIG. 16, the lane position 107 estimated on the radio wave image is displayed superimposed on the visible image. FIG. 16 also shows a boundary point 109 that is a result of estimating the lane boundary by image analysis. As a result, the position of the lane can be estimated and displayed even in the low contrast region 103.

  When the position of the lane is thus displayed, the lane estimation process according to the present embodiment is terminated. Thereafter, the image generation unit 11 outputs an output image on which the position of the lane is displayed.

[Effect of the third embodiment]
As described above in detail, the image generation device 21 according to the present embodiment further includes the lane estimation unit 23 that sets the lane detection region on the visible image and estimates the position of the lane. Then, when the lane detection region is set at least partially in the low-contrast region of the visible image, the image generation unit 11 generates an output image in which the contrast of the region including at least the lane detection region is improved. The estimation unit 23 estimates a lane position by setting a lane detection region on the output image. Thereby, even when there is a low-contrast region on the visible image due to nighttime or backlighting, the position of the lane can be estimated and displayed.

  The above-described embodiment is an example of the present invention. For this reason, the present invention is not limited to the above-described embodiment, and even if it is a form other than this embodiment, as long as it does not depart from the technical idea of the present invention, it depends on the design and the like. Of course, various modifications are possible.

DESCRIPTION OF SYMBOLS 1, 21 Image generation apparatus 3 Radio wave reception part 5 Visible image imaging part 7 Object classification part 9 Contrast detection part 11 Image generation part 23 Lane estimation part 31 Sky 33 Road surface 35 Rough ground 37, 71 Road surface area 39, 73, 103 Low contrast Area 41 Area 75 corresponding to road area among low contrast areas High contrast area 81, 83 Solid object detection area 85, 87 Area 101, 105 corresponding to solid object detection area Lane detection area 107 Lane position 109 Boundary point

Claims (5)

An image generation device that generates an image from a radio wave image generated by detecting electromagnetic waves coming from the front of the vehicle and a visible image obtained by imaging the front of the vehicle,
A radio wave receiving unit that generates a radio wave image by receiving a horizontal polarization component and a vertical polarization component of electromagnetic waves coming from the front of the vehicle;
A visible image capturing unit that captures a visible image in front of the vehicle;
An object classification unit for classifying an object on a radio wave image generated by the radio wave reception unit based on the horizontal polarization component and the vertical polarization component, and setting the classified region as an object region;
A contrast detection unit that detects a low-contrast region where contrast is reduced on the visible image captured by the visible image capturing unit;
An image generation unit that generates an output image in which a contrast of a region corresponding to an object region on the radio wave image among the low contrast regions on the visible image is improved with respect to the visible image. An image generating device. The radio wave reception unit generates a high resolution radio wave image for the object region,
The said image generation part synthesize | combines the said radio wave image to the area | region corresponding to the object area | region on the said radio wave image among the low contrast area | regions on the said visible image, The said output image is produced | generated. The image generating apparatus described. The visible image capturing unit captures a brightness-enhanced visible image in which the brightness of a region corresponding to the object region on the radio wave image is increased among low-contrast regions on the visible image;
The image generation apparatus according to claim 1, wherein the image generation unit generates the brightness-increased visible image as the output image. Further comprising a lane estimation unit that sets a lane detection region on the visible image captured by the visible image capturing unit and estimates a lane position from an image analysis result of the lane detection region;
The image generation unit generates an output image in which the contrast of at least the area including the lane detection area is improved when the lane detection area is at least partially set in the low contrast area on the visible image. ,
The image generation apparatus according to claim 1, wherein the lane estimation unit estimates the position of a lane by setting the lane detection region on the output image. An image generation method by an image generation device that generates an image from a radio wave image generated by detecting electromagnetic waves coming from the front of a vehicle and a visible image obtained by imaging the front of the vehicle,
The image generation device includes:
Receiving a horizontal polarization component and a vertical polarization component of electromagnetic waves coming from the front of the vehicle to generate a radio wave image;
Taking a visible image in front of the vehicle,
Classifying the object on the generated radio wave image based on the horizontal polarization component and the vertical polarization component and setting the classified region as the object region,
Detecting a low-contrast region where the contrast is reduced on the captured visible image;
An image generation method, comprising: generating an output image in which a contrast of a region corresponding to an object region on the radio wave image among the low contrast regions on the visible image is improved with respect to the visible image.