Image processing apparatus, control apparatus, image processing method, and program

Abstract

To provide an image processing apparatus, a control apparatus, an image processing method, and a program capable of easily estimating whether or not highly accurate image recognition can be performed.SOLUTION: An image processing apparatus includes: a first acquisition unit that acquires information on an optical system included in an image pickup apparatus configured to capture an image; a second acquisition unit that acquires, on the basis of the information on the optical system, a change amount in the information on the optical system in a peripheral area of a target pixel in the image; and a third acquisition unit that acquires, on the basis of the change amount, information on recognition accuracy in the peripheral area by a recognition apparatus configured to perform image recognition.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an image processing apparatus, a control apparatus, an image processing method, and a program.

Background Art

[0002] Conventionally, there is known a configuration in which an image in which the type of an object is known is used, and a classifier that has learned the relationship between the type of the object and the features on the image is used to recognize the type of the object shown in the captured image. In the learning, an image with little distortion or an image with corrected distortion is used. However, when recognizing the type of an object shown in an image acquired using an optical system with a large distortion, since the object has a shape different from the learned shape, the recognition accuracy is low. Patent Document 1 discloses a configuration in which recognition is performed using an image after correcting distortion. Further, Patent Document 2 discloses a configuration in which learning is performed using an image having distortion.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the configuration of Patent Document 1, since distortion correction is performed for each captured image, the processing time becomes long and resources of the arithmetic unit are required. Further, in the configuration of Patent Document 2, since learning is performed for each optical system, the network of the classifier tends to become large.

[0005] An object of the present invention is to provide an image processing apparatus, a control apparatus, an image processing method, and a program capable of estimating whether high-precision image recognition can be simply executed.

Means for Solving the Problem

[0006] An image processing apparatus according to one aspect of the present invention includes a first acquisition unit that acquires information regarding an optical system included in an imaging apparatus that captures an image, a second acquisition unit that acquires a change amount of information regarding the optical system in a peripheral region of a target pixel in the image based on the information regarding the optical system, and a third acquisition unit that acquires information regarding recognition accuracy in the peripheral region by a recognition apparatus that performs image recognition based on the change amount.

Advantages of the Invention

[0007] According to the present invention, it is possible to provide an image processing apparatus, a control apparatus, an image processing method, and a program capable of estimating whether high-precision image recognition can be easily executed.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Embodiments for Carrying Out the Invention

[0009] Hereinafter, embodiments and examples of the present invention will be described in detail with reference to the drawings. In each figure, the same members are denoted by the same reference numerals, and redundant descriptions are omitted.

[0010] FIG. 1 is a diagram showing the configuration of an image recognition system 1 according to an embodiment of the present invention. The image recognition system 1 includes an imaging device 10 and a control device 20. The imaging device 10 and the control device 20 are communicably connected to each other. In the present embodiment, the imaging device 10 and the control device 20 are configured as separate bodies, but the imaging device 10 may be configured to have the functions of the control device 20.

[0011] The imaging device 10 includes an imaging optical system 11 which is an optical member composed of a lens or the like, and an image sensor 12 that captures an image formed by the imaging optical system 11, and captures an image by photographing the periphery of the installation location. The imaging optical system 11 preferably uses a lens with a wide angle of view. The imaging device 10 transmits the acquired image to the control device 20.

[0012] The control device 20 is composed of a computer (for example, ECU: Electronic Control Unit) including at least one processor, a memory, etc., and performs arithmetic processing such as image processing using the image acquired from the imaging device 10. The control device 20 may have a function of controlling the imaging device 10 to perform imaging.

[0013] The control device 20 includes a recognition device 21 and an image processing device 22. The recognition device 21 performs image recognition processing on the image captured by the imaging device 10, and detects an object (subject) included in the image and recognizes the type of the object. The recognition device 21 includes, for example, a classifier that has learned the relationship between the type of the object and the features on the image. In this way, the recognition device 21 can perform image recognition on the captured image and recognize the type of the object shown in the image. Further, the recognition device 21 can perform real-time image recognition by repeatedly performing recognition processing using a moving image or a series of images.

[0014] The image processing device 22 includes a first acquisition unit 22a, a second acquisition unit 22b, a third acquisition unit 22c, and a control unit 22d.

[0015] The first acquisition unit 22a acquires information about the imaging optical system 11. The information about the imaging optical system 11 is information indicating the optical characteristics or features of the imaging optical system 11. As a more specific example of the information about the imaging optical system 11, it is information about the image magnification that affects the magnitude of the distortion aberration.

[0016] FIG. 2 is a diagram showing an example of a parameter representing the image magnification of an image acquired using the imaging optical system 11. The larger the value of the parameter, the larger the image magnification. FIGS. 2(a) and 2(b) respectively show parameters representing the image magnification in the X direction and the Y direction. As shown in FIGS. 2(a) and 2(b), the imaging optical system 11 is configured such that the image magnification is larger for the pixels arranged in the center.

[0017] FIG. 3 is a diagram showing an example of the imaging optical system 11. In FIG. 3, the left side is the magnified conjugate side (object side), and the right side is the reduced conjugate side (image side). The imaging optical system 11 is an imaging (image-forming) optical system that condenses a light beam from an object (not shown) located on the magnified conjugate side and forms an object image on the image plane 300 on the reduced conjugate side.

[0018] The imaging optical system 11 has a front group 111 including a plurality of lenses, an aperture stop ST, and a rear group 112 including a plurality of lenses, which are arranged in order from the magnified conjugate side to the reduced conjugate side. An IR cut filter 201 and a cover glass 202 are arranged between the imaging optical system 11 and the image plane 300. Also, a low-pass filter or the like may be additionally arranged as necessary, or the IR cut filter 201 or the like may be omitted.

[0019] The front group 111 includes lenses L1, L2, L3, and L4 arranged in order from the enlarged conjugate side to the reduced conjugate side. Also, the rear group 112 includes L5, L6, L7, and L8 arranged in order from the enlarged conjugate side to the reduced conjugate side. The lens L1 arranged on the most enlarged conjugate side in the front group 111 is an aspherical lens (first aspherical lens) in which both sides on the enlarged conjugate side and the reduced conjugate side are formed as aspherical surfaces, and the paraxial refractive power (paraxial power) is negative. The lens L2 arranged second from the enlarged conjugate side in the front group 111 is an aspherical lens (second aspherical lens) in which both sides are formed as aspherical surfaces, and the paraxial refractive power is positive. The lenses L3 and L4 arranged third and fourth from the enlarged conjugate side in the front group 111 are both spherical lenses, and their respective refractive powers are negative and positive. The lenses L5, L6, and L7 arranged on the most enlarged conjugate side, second, and third from the enlarged conjugate side in the rear group 112 are all spherical lenses, and their respective refractive powers are negative, positive, and negative. The lens (final lens) L8 arranged on the most reduced conjugate side in the rear group 112 is an aspherical lens (third aspherical lens) in which both sides are formed as aspherical surfaces, and the paraxial refractive power is positive.

[0020] The imaging optical system 11 is an optical system in which the angle formed by the optical axis and the outermost chief ray, that is, the maximum half field angle θmax is π / 2 (= 90°), and has a maximum half field angle equivalent to that of a fish-eye lens. However, the value of the maximum half field angle θmax is an example, and the present invention is not limited thereto. Also, the imaging optical system 11 is an optical system having a larger imaging magnification in the angular field region near the center (hereinafter referred to as the central angular field region) compared to a fish-eye lens.

[0021] FIG. 4 is a diagram showing another example of the optical characteristics of the imaging optical system 11. FIGS. 4(a) and 4(b) show the projection characteristics and the resolution characteristics of the imaging optical system 11, respectively. In FIG. 4, °(deg) is used as the unit of the angle of view. The projection characteristic y(θ) shown in FIG. 4(a) represents the relationship between the half angle of view (the angle between the optical axis and the incident light ray) θ and the imaging height (image height) y on the image plane 300. Also, the resolution characteristic shown in FIG. 4(b) represents the amount of change in the imaging height y with respect to a minute change in the angle of view at the half angle of view θ, that is, the differential value dy(θ) / dθ of the projection characteristic y(θ) at the half angle of view θ. The differential value dy(θ) / dθ corresponds to the local resolution at the imaging height y. The larger the value of the differential value dy(θ) / dθ, the higher the local resolution. A high local resolution means a large local imaging magnification.

[0022] The imaging optical system 11 has a higher central resolution than that of an orthographic projection system (y(θ)=f×sinθ) with a high resolution in the central angle of view region (hereinafter referred to as the central resolution) among the projection methods of a general fisheye lens.

[0023] Also, by suppressing the distortion associated with the optical characteristics of the imaging optical system 11 in the low angle of view region (central angle of view region), the distortion near the center of the image is reduced. Therefore, a natural perspective can be obtained visually for the image in the low angle of view region. That is, it is possible to eliminate the need for performing an electronic distortion correction process or to reduce the correction amount.

[0024] Since the imaging optical system 11 has a wider region (angle of view) in which the distortion associated with the optical characteristics of the imaging optical system 11 is suppressed than a general fisheye lens, it is possible to make the region where the above-described electronic distortion correction process is not required wider than that of a general fisheye lens.

[0025] Incidentally, the imaging optical system 11 may be configured to have a higher central resolution than that of the orthographic projection method, which has a high central resolution among the projection methods of general fish-eye lenses as described above, and to suppress distortion associated with optical characteristics in the low angle-of-view region, but the present invention is not limited thereto. For example, the imaging optical system 11 may be configured to have reverse characteristics.

[0026] Specifically, the imaging optical system 11 may be configured to have a projection characteristic in which the resolution in the high angle-of-view region (peripheral angle-of-view region) is higher than the resolution in the central angle-of-view region. In other words, the imaging optical system 11 may be configured to have a projection characteristic in which the increase rate (slope) of the image height y is small in the central region near the optical axis where the angle of view is small, and the increase rate of the image height y increases as the angle of view increases in the peripheral region. In this case, the projection characteristic preferably has a larger change in the image height y than that of a general equidistant projection (y = fθ), and further than that of a stereographic projection (y = 2ftan(θ / 2)). Since the imaging optical system 11 configured in this way also has a wider region (angle of view) in which distortion associated with the optical characteristics of the imaging optical system 11 is suppressed than a general fish-eye lens, it is possible to take a wider region than a general fish-eye lens in which the above-described electronic distortion correction process is not required.

[0027] The second acquisition unit 22b acquires the amount of change in information regarding the imaging optical system 11 in the peripheral region (frame region) of the pixel of interest in the image acquired by the imaging device 10. For example, when the peripheral regions 101 and 102 are set, the second acquisition unit 22b acquires the amount of change in the peripheral regions 101 and 102. In the present embodiment, the second acquisition unit 22b acquires the amount of change in information regarding the imaging optical system 11 in the peripheral region using a differential filter, but other methods may be used.

[0028] The third acquisition unit 22c acquires (calculates) information regarding the recognition accuracy by the recognition device 21 in the peripheral area based on the amount of change acquired by the second acquisition unit 22b. In the present embodiment, the third acquisition unit 22c acquires information regarding the recognition accuracy by the recognition device 21 in the peripheral area by using information indicating the correspondence relationship between the amount of change in the information regarding the imaging optical system 11 and the recognition accuracy by the recognition device 21. For example, the third acquisition unit 22c acquires information regarding the recognition accuracy in the peripheral area by the recognition device 21 by using a table showing the correspondence relationship between the amount of change in the focal length shown in FIG. 5 and the recognition accuracy by the recognition device 21. The information regarding the recognition accuracy is information representing the recognition accuracy, and may be information representing the recognition accuracy as high or low as shown in FIG. 5, or may be information represented by a numerical value.

[0029] Here, as shown in FIG. 2, the amount of change in the information regarding the imaging optical system 11 (parameters indicating the image magnification) in the peripheral area 101 is small, and the amount of change in the information regarding the imaging optical system 11 in the peripheral area 102 is large. Therefore, the third acquisition unit 22c acquires the recognition accuracy such that the recognition accuracy of the target pixel corresponding to the peripheral area 101 is higher than the recognition accuracy of the target pixel corresponding to the peripheral area 102.

[0030] Note that when the third acquisition unit 22c acquires information regarding the recognition accuracy by the recognition device 21 in the peripheral area, in the present embodiment, a table showing the correspondence relationship between the amount of change in the focal length and the recognition accuracy by the recognition device 21 is used, but the present invention is not limited to this. For example, the third acquisition unit 22c may use an expression showing the correspondence relationship between the amount of change in the focal length and the recognition accuracy by the recognition device 21.

[0031] The control unit 22d performs processing by using the information regarding the recognition accuracy by the recognition device 21 in the peripheral area acquired by the third acquisition unit 22c.

[0032] By having the above configuration, the image processing device 22 can estimate whether it can easily execute high-precision image recognition.

[0033] Hereinafter, the method of image recognition by the image recognition system 1 of each embodiment will be described.

Embodiment

[0034] In this embodiment, the control unit 22d determines the target pixel used for image recognition by the recognition device 21. Specifically, when the information regarding the recognition accuracy by the recognition device 21 in the peripheral area acquired by the third acquisition unit 22c satisfies a predetermined condition, the control unit 22d determines to cause the recognition device 21 to use the target pixel corresponding to the peripheral area. On the other hand, when the information regarding the recognition accuracy in the peripheral area does not satisfy the predetermined condition, the control unit 22d determines not to cause the recognition device 21 to use the target pixel corresponding to the peripheral area.

[0035] Note that, in this embodiment, no predetermined image correction process (distortion correction) is performed on the target pixel determined by the control unit 22d to be used by the recognition device 21. That is, in this embodiment, when the information regarding the recognition accuracy by the recognition device 21 in the peripheral area satisfies a predetermined condition, the control unit 22d determines to cause the recognition device 21 to use the target pixel corresponding to the peripheral area without performing a predetermined image correction process (distortion correction).

[0036] Note that the control unit 22d may determine the peripheral area used for image recognition by the recognition device 21. Specifically, when the information regarding the recognition accuracy by the recognition device 21 in the peripheral area acquired by the third acquisition unit 22c satisfies a predetermined condition, the control unit 22d may determine to cause the recognition device 21 to use the peripheral area. In this case, when the information regarding the recognition accuracy in the peripheral area does not satisfy the predetermined condition, the control unit 22d determines not to cause the recognition device 21 to use the peripheral area.

[0037] FIG. 6 is a flowchart showing the method of image recognition of the image recognition system 1 of this embodiment.

[0038] In step S101, the first acquisition unit 22a acquires information regarding the imaging optical system 11.

[0039] In step S102, the second acquisition unit 22b acquires the amount of change in information regarding the imaging optical system 11 in the peripheral region of the target pixel in the image acquired by the imaging device 10. In this embodiment, the peripheral region is set to have a predetermined size. For example, the peripheral region may be set to have the same size at all the viewing angles of the imaging device 10, that is, the number of pixels included in the peripheral region may be set to a fixed value. Further, the peripheral region may be set according to the viewing angle of the imaging device 10, for example, according to the expected value of the size of the object at each viewing angle.

[0040] In step S103, the third acquisition unit 22c acquires information regarding the recognition accuracy by the recognition device 21 in the peripheral region based on the amount of change acquired by the second acquisition unit 22b in step S102.

[0041] In step S104, the image processing device 22 determines the target pixel corresponding to the peripheral region used for image recognition by the recognition device 21. Specifically, the image processing device 22 first determines whether the information regarding the recognition accuracy by the recognition device 21 in the peripheral region acquired by the third acquisition unit 22c in step S103 satisfies a predetermined condition. For example, when the information regarding the recognition accuracy by the recognition device 21 in the peripheral region is represented by a numerical value so as to increase in proportion to the highness of the recognition accuracy, it may be determined whether the information regarding the recognition accuracy by the recognition device 21 in the peripheral region is equal to or greater than a predetermined value. Next, when the image processing device 22 determines that the information regarding the recognition accuracy by the recognition device 21 in the peripheral region satisfies the predetermined condition, it determines to cause the recognition device 21 to use the target pixel corresponding to the peripheral region. On the other hand, when the image processing device 22 determines that the information regarding the recognition accuracy by the recognition device 21 in the peripheral region does not satisfy the predetermined condition, it determines not to cause the recognition device 21 to use the target pixel corresponding to the peripheral region.

[0042] Note that the control unit 22d may cut out from the image acquired by the imaging device 10 a region (second region) including a region (first region) composed of the target pixels determined to be used by the recognition device 21. The second region may be the same size as the first region or may be larger than the first region. By making the second region larger than the first region, it is possible to suppress the recognition device 21 from being unable to perform accurate image recognition at the cutout boundary. Further, when the first region is a peripheral portion of the image acquired by the imaging device 10, the image processing device 22 may divide and cut out the second region.

[0043] Further, the control unit 22d may fill in a region (fourth region) including at least a part of a region (third region) composed of the target pixels determined not to be used by the recognition device 21 in the image acquired by the imaging device 10. The fourth region may be the same size as the third region or may be smaller than the third region. By making the fourth region smaller than the third region, it is possible to suppress the recognition device 21 from being unable to perform accurate image recognition at the cutout boundary.

[0044] In step S105, the recognition device 21 performs image recognition on the target pixels determined by the control unit 22d to be used by the recognition device 21 in step S104, and recognizes the type of the object shown in the image.

[0045] According to the configuration of the present embodiment, since the number of pixels on which the recognition device 21 performs image recognition can be suppressed, the processing time can be shortened.

Embodiment

[0046] In the present embodiment, the control unit 22d associates the result of image recognition performed using the image acquired by the imaging device 10 by the recognition device 21 with the information regarding the recognition accuracy by the recognition device 21 in the peripheral region acquired by the third acquisition unit 22c.

[0047] FIG. 7 is a flowchart showing a method of image recognition of the image recognition system 1 of the present embodiment. It is assumed that while the flow of FIG. 7 is being executed (until the process of step S204 is executed), the recognition device 21 performs image recognition using the image acquired by the imaging device 10 and recognizes the type of object shown in the image.

[0048] In step S201, the first acquisition unit 22a acquires information regarding the imaging optical system 11.

[0049] In step S202, the second acquisition unit 22b acquires the amount of change in the information regarding the imaging optical system 11 in the peripheral region of the target pixel in the image acquired by the imaging device 10. In the present embodiment, the peripheral region is set to have a predetermined size. For example, the peripheral region may be set to have the same size at all the viewing angles of the imaging device 10, that is, the number of pixels included in the peripheral region may be set to a fixed value. Further, the peripheral region may be set according to the viewing angle of the imaging device 10, for example, according to the expected value of the size of the object at each viewing angle. In the present embodiment, the second acquisition unit 22b acquires the amount of change in the information regarding the imaging optical system 11 over the entire region of the image.

[0050] In step S203, the third acquisition unit 22c acquires information regarding the recognition accuracy by the recognition device 21 in the peripheral region based on the amount of change acquired by the second acquisition unit 22b in step S202.

[0051] In step S204, the control unit 22d associates the result of image recognition performed using the image acquired by the imaging device 10 by the recognition device 21 with the information regarding the recognition accuracy by the recognition device 21 in the peripheral area acquired by the third acquisition unit 22c in step S203. Then, in this embodiment, the control unit 22d determines the information to be used among the results of image recognition by the recognition device 21. Specifically, the control unit 22d first determines whether the information regarding the recognition accuracy by the recognition device 21 in the peripheral area satisfies a predetermined condition. For example, when the information regarding the recognition accuracy by the recognition device 21 in the peripheral area is represented by a numerical value such that it increases in proportion to the highness of the recognition accuracy, it may be determined whether the information regarding the recognition accuracy by the recognition device 21 in the peripheral area is equal to or greater than a predetermined value. Next, when the control unit 22d determines that the information regarding the recognition accuracy by the recognition device 21 in the peripheral area satisfies the predetermined condition, it determines to use the result of image recognition by the recognition device 21 performed using the peripheral area. On the other hand, when the control unit 22d determines that the information regarding the recognition accuracy by the recognition device 21 in the peripheral area does not satisfy the predetermined condition, it determines not to use the result of image recognition by the recognition device 21 performed using the peripheral area.

[0052] Note that the control unit 22d may superimpose the information regarding the recognition accuracy by the recognition device 21 in the peripheral area on the result of image recognition performed using the image acquired by the imaging device 10 by the recognition device 21.

[0053] In addition, the control unit 22d may hold not only the information associating the current image recognition result by the recognition device 21 with the information regarding the recognition accuracy by the recognition device 21 in the peripheral area, but also the information associated in the past. That is, the control unit 22d may hold the information associating the image recognition result by the recognition device 21 and the information regarding the recognition accuracy by the recognition device 21 in the peripheral area, which are acquired at the first timing and the second timing after the first timing. The control unit 22d may determine to use the image recognition result by the recognition device 21 performed using the peripheral area where the information regarding the recognition accuracy by the recognition device 21 does not satisfy a predetermined condition, using the information associated at different timings. For example, the image recognition result performed using the peripheral area determined not to satisfy the predetermined condition after being determined to satisfy the predetermined condition regarding the information regarding the recognition accuracy by the recognition device 21 in the peripheral area may be used. Further, even in the case of a peripheral area determined a plurality of times not to satisfy the predetermined condition regarding the information regarding the recognition accuracy by the recognition device 21 in the peripheral area, if all the image recognition results performed using the peripheral area are the same, the results may be used.

[0054] According to the configuration of this embodiment, by associating the image recognition result performed using the image acquired by the imaging device 10 with the information regarding the recognition accuracy by the recognition device 21 in the peripheral area, it is possible to perform higher-precision image recognition as compared with the first embodiment.

Embodiment

[0055] In this embodiment, a case where the peripheral area is set based on the image recognition result performed using the image acquired by the imaging device 10 by the recognition device 21 will be described. In this embodiment, the control unit 22d determines the information to be used among the image recognition results by the recognition device 21.

[0056] FIG. 8 is a flowchart showing a method of image recognition of the image recognition system 1 of the present embodiment. While the flow of FIG. 8 is being executed (until the process of step S302 is executed), it is assumed that the recognition device 21 performs image recognition using the image acquired by the imaging device 10 and recognizes the type of object shown in the image.

[0057] In step S301, the first acquisition unit 22a acquires information regarding the imaging optical system 11.

[0058] In step S302, the second acquisition unit 22b acquires a change amount of information regarding the imaging optical system 11 in a peripheral region of a target pixel in the image acquired by the imaging device 10. In the present embodiment, the peripheral region is set based on the result of image recognition performed using the image acquired by the imaging device 10.

[0059] In step S303, the third acquisition unit 22c acquires information regarding the recognition accuracy by the recognition device 21 in the peripheral region based on the change amount acquired by the second acquisition unit 22b in step S102.

[0060] In step S304, the control unit 22d determines information to be used among the results of image recognition by the recognition device 21. Specifically, the control unit 22d first determines whether the information regarding the recognition accuracy by the recognition device 21 in the peripheral region acquired by the third acquisition unit 22c in step S203 satisfies a predetermined condition. For example, when the information regarding the recognition accuracy by the recognition device 21 in the peripheral region is represented by a numerical value so as to increase in proportion to the highness of the recognition accuracy, it may be determined whether the information regarding the recognition accuracy by the recognition device 21 in the peripheral region is equal to or greater than a predetermined value. Next, when the control unit 22d determines that the information regarding the recognition accuracy by the recognition device 21 in the peripheral region satisfies the predetermined condition, the control unit 22d determines to use the result of image recognition performed using the peripheral region. On the other hand, when the control unit 22d determines that the information regarding the recognition accuracy by the recognition device 21 in the peripheral region does not satisfy the predetermined condition, the control unit 22d determines not to use the result of image recognition performed using the peripheral region.

[0061] According to the configuration of this embodiment, it is possible to recognize objects in a wider range compared to Embodiment 1. [Other Embodiments] The present invention can also be realized by supplying a program that realizes one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and causing one or more processors in a computer of the system or apparatus to read and execute the program. It can also be realized by a circuit (for example, an ASIC) that realizes one or more functions.

[0062] The disclosure of this embodiment includes the following configurations and methods.

[0063] (Configuration 1) A first acquisition unit that acquires information regarding an optical system included in an imaging device that images an image, A second acquisition unit that acquires a change amount of the information in a peripheral region of a pixel of interest in the image, An image processing apparatus, comprising: a third acquisition unit that acquires information regarding recognition accuracy in the peripheral region by a recognition device that performs image recognition using the image based on the change amount. (Configuration 2) The image processing apparatus according to Configuration 1, wherein the third acquisition unit acquires information regarding the recognition accuracy using information indicating a correspondence relationship between the change amount and the recognition accuracy. (Configuration 3) The image processing apparatus according to Configuration 1 or 2, further comprising a control unit that determines to cause the recognition device to use the peripheral region when the information regarding the recognition accuracy satisfies a predetermined condition, and determines not to cause the recognition device to use the peripheral region when the information regarding the recognition accuracy does not satisfy the predetermined condition. (Configuration 4) The image processing apparatus according to Configuration 3, wherein the control unit determines to cause the recognition device to use the pixel of interest without performing a predetermined image correction process when the information regarding the recognition accuracy satisfies a predetermined condition. (Configuration 5) The image processing apparatus according to Configuration 3, wherein the control unit cuts out a second region including a first region composed of the peripheral region determined to be used by the recognition device from the image. (Configuration 6) The image processing apparatus according to Configuration 3, wherein the control unit fills in a fourth region including at least a part of a third region composed of the peripheral region determined to be used by the recognition device in the image. (Configuration 7) The image processing apparatus according to Configuration 1 or 2, further comprising a control unit that associates the result of image recognition performed using the image by the recognition device with information regarding the recognition accuracy. (Configuration 8) The image processing apparatus according to Configuration 7, wherein when the information regarding the recognition accuracy satisfies a predetermined condition, the control unit determines to use the result of image recognition performed using the peripheral region by the recognition device, and when the information regarding the recognition accuracy does not satisfy the predetermined condition, the control unit determines not to use the result of image recognition performed using the peripheral region by the recognition device. (Configuration 9) The image processing apparatus according to Configuration 7 or 8, wherein the control unit superimposes the information regarding the recognition accuracy on the result of image recognition performed using the image by the recognition device. (Configuration 10) The image processing apparatus according to any one of Configurations 7 to 9, wherein the control unit holds information associating the result of image recognition acquired at a first timing with the information regarding the recognition accuracy, and information associating the result of image recognition acquired at a second timing after the first timing with the information regarding the recognition accuracy. (Configuration 11) The image processing apparatus according to any one of Configurations 1 to 10, wherein the number of pixels included in the peripheral region is a fixed value. (Configuration 12) The image processing apparatus according to any one of Configurations 1 to 10, wherein the number of pixels included in the peripheral region is set according to the angle of view of the imaging device. (Configuration 13) The image processing apparatus according to any one of Configurations 1 to 6, wherein the peripheral area is set based on the result of image recognition by the recognition apparatus performed using the image. (Configuration 14) An image processing apparatus according to any one of Configurations 1 to 15, and a control apparatus comprising a recognition apparatus that performs image recognition. (Method 1) A step of acquiring information regarding an optical system included in an imaging apparatus that images an image; a step of acquiring a change amount of information regarding the optical system in a peripheral area of a pixel of interest in the image based on the information regarding the optical system; and a step of acquiring information regarding recognition accuracy in the peripheral area by a recognition apparatus that performs image recognition based on the change amount. An image processing method characterized by comprising the steps. (Configuration 15) A program characterized by causing a computer to execute the image processing method according to Method 1.

[0064] As described above, the preferred embodiments of the present invention have been described. However, the present invention is not limited to these embodiments, and various modifications and changes are possible within the scope of the gist thereof.

Explanation of Reference Numerals

[0065] 10 Imaging apparatus 12 Imaging optical system (optical system) 22 Image processing apparatus 22a First acquisition unit 22b Second acquisition unit 22c Third acquisition unit

Claims

1. A first acquisition unit that acquires information regarding an optical system included in an imaging device that captures an image; A second acquisition unit that acquires a change amount of information regarding the optical system in a peripheral region of a pixel of interest in the image based on the information regarding the optical system; A third acquisition unit that acquires information regarding recognition accuracy in the peripheral region by a recognition device that performs image recognition based on the change amount; An image processing apparatus, comprising:

2. The image processing apparatus according to claim 1, wherein the third acquisition unit acquires information regarding the recognition accuracy by using information indicating a correspondence relationship between the change amount and the recognition accuracy.

3. The image processing apparatus according to claim 1 or 2, further comprising a control unit that determines to cause the recognition device to use the pixel of interest when the information regarding the recognition accuracy satisfies a predetermined condition, and determines not to cause the recognition device to use the pixel of interest when the information regarding the recognition accuracy does not satisfy the predetermined condition.

4. The image processing apparatus according to claim 3, wherein the control unit determines to cause the recognition device to use the pixel of interest without performing predetermined image correction processing when the information regarding the recognition accuracy satisfies a predetermined condition.

5. The image processing apparatus according to claim 3, wherein the control unit cuts out a second region including a first region composed of the pixels of interest determined to be used by the recognition device from the image.

6. The image processing apparatus according to claim 3, wherein the control unit fills in a fourth region including at least a part of a third region composed of the pixels of interest determined not to be used by the recognition device in the image.

7. The image processing apparatus according to claim 1 or 2, further comprising a control unit that associates a result of image recognition performed using the image by the recognition device with the information regarding the recognition accuracy.

8. When the information regarding the recognition accuracy satisfies a predetermined condition, the control unit determines to use the result of the image recognition performed by the recognition device using the peripheral area, and when the information regarding the recognition accuracy does not satisfy the predetermined condition, the control unit determines not to use the result of the image recognition performed by the recognition device using the peripheral area. The image processing apparatus according to claim 7, characterized in that.

9. The control unit superimposes the information regarding the recognition accuracy on the result of the image recognition performed by the recognition device using the image. The image processing apparatus according to claim 7, characterized in that.

10. The control unit holds information associating the result of the image recognition acquired at the first timing with the information regarding the recognition accuracy, and information associating the result of the image recognition acquired at the second timing after the first timing with the information regarding the recognition accuracy. The image processing apparatus according to claim 7, characterized in that.

11. The number of pixels included in the peripheral area is a fixed value. The image processing apparatus according to claim 1 or 2, characterized in that.

12. The number of pixels included in the peripheral area is set according to the angle of view of the imaging device. The image processing apparatus according to claim 1 or 2, characterized in that.

13. The peripheral area is set based on the result of the image recognition by the recognition device performed using the image. The image processing apparatus according to claim 1 or 2, characterized in that.

14. The image processing apparatus according to claim 1 or 2, and A recognition device that performs image recognition, and A control device, characterized in that it has.

15. A step of acquiring information regarding an optical system included in an imaging device that images an image, and Based on the information regarding the optical system, obtaining a change amount of the information regarding the optical system in a peripheral region of a pixel of interest in the image; Based on the change amount, obtaining information regarding recognition accuracy in the peripheral region by a recognition device that performs image recognition; An image processing method, characterized by comprising the above steps.

16. A program, characterized by causing a computer to execute the image processing method according to claim 15.

Images