Recognition device, and image correction program

Abstract

To provide a recognition device, and an image correction program which do not require adjustment of an imaging device and reconstruction of a system.SOLUTION: A recognition device comprises a reception unit and a control unit. The reception unit receives a first image having environment information related to imaging environment and a second image having the environment information related to the imaging environment different from that of the first image. The control unit calculates the environment information related to the imaging environment from the first image, creates a filter for correcting the second image, and performs control which executes correction based on the filter with respect to the second image.SELECTED DRAWING: Figure 7

Description

[Technical field]

[0001] The present invention relates to a recognition device and an image correction program. [Background technology]

[0002] Conventionally, there are recognition devices that recognize objects and human behavior. Some of these recognition devices are connected to multiple image capture devices. In a system in which such recognition devices and image capture devices are connected, adjustments are made to each image capture device and the system is reconstructed so that the image quality is not affected by the installation environment of each image capture device. However, it is time-consuming to perform adjustments and system reconstruction for each image capture device every time a new image capture device is connected or the installation environment changes. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2016-015116 A Summary of the Invention [Problem to be solved by the invention]

[0004] An object of the present invention is to provide a recognition device and an image correction program that eliminate the need to adjust an imaging device or reconstruct a system. [Means for solving the problem]

[0005] In order to achieve the above object, the recognition device of this embodiment includes a receiving unit and a control unit. The receiving unit receives a first image and a second image. The control unit calculates environmental information related to an imaging environment from the first image, creates a filter for correcting the second image, and performs control to perform correction on the second image based on the filter. [Brief description of the drawings]

[0006] [Figure 1]FIG. 1 is a block diagram showing an example of an image recognition system according to a first embodiment; [Diagram 2] FIG. 1 is a block diagram illustrating an example of a recognition device according to a first embodiment. [Diagram 3] FIG. 1 is a block diagram showing an example of an imaging device according to a first embodiment; [Figure 4] FIG. 1 is a schematic diagram showing an example of an environment information table according to a first embodiment; [Diagram 5] 1 is a flowchart showing an example of an environment information registration process of an imaging device according to a first embodiment. [Figure 6] 1 is a flowchart showing an example of an environment information registration process of the recognition device according to the first embodiment. [Figure 7] 1 is a flowchart showing an example of a filtering process according to the first embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0007] (First embodiment) Hereinafter, the present embodiment will be described with reference to the drawings. FIG. 1 is a block diagram showing an example of an image recognition system 1 according to a first embodiment. The image recognition system 1 is composed of, for example, a recognition device 10 and an imaging device, and each device is connected via a wired or wireless LAN (Local Area Network). The recognition device 10 and the imaging device 20 constituting the image recognition system 1 may be installed in the same building or in different buildings.

[0008] The recognition device 10 executes image recognition. The recognition device 10 recognizes objects, human behavior, etc. from images received from external devices such as imaging devices. The recognition device 10 also represents the environment in which each imaging device 20 is installed using brightness, frequency, color temperature, etc., and stores it as environmental information. When recognizing objects, human behavior, etc. from an image, the recognition device 10 also executes a filtering process on the image. In this embodiment, the filtering process refers to a process of correcting environmental information such as brightness, frequency, and color temperature of a received image so that the brightness, frequency, and color temperature become the same as the environmental information registered in advance.

[0009] The imaging device 20 captures an image of a preset imaging range. The imaging device 20 outputs light focused on an imaging element as an electrical signal, and generates image data by converting, compressing, and encoding the signal. After generating the image data, the imaging device 20 transmits the image data to the recognition device 10. In addition, the imaging device 20 performs development processing including tone conversion processing, noise removal, and scratch correction when generating the image data. Furthermore, the imaging device 20 may start imaging when communication with the recognition device 10 is established.

[0010] Each imaging device 20 stores a unique ID (hereinafter, device ID) for identifying the imaging device, and transmits the ID to the recognition device 10 together with the image data when the image data is transmitted.

[0011] The process of registering environmental information in the recognition device 10 may be performed automatically when the imaging device 20 is connected, or may be performed when switching modes by implementing a dedicated mode in the imaging device 20 for capturing images that will become environmental information.

[0012] 2 is a block diagram showing an example of the recognition device 10 according to the first embodiment. The recognition device 10 is composed of, for example, a control unit 100, a storage unit 101, and a communication I / F 102, and each unit is connected via a bus.

[0013] The control unit 100 includes a central processing unit (CPU) 1001, a read only memory (ROM) 1002, and a random access memory (RAM) 1003. The CPU 1001 controls the entire recognition device 10. The ROM 1002 stores various programs, such as a program used to drive the CPU 1001, and various data. The RAM 1003 is used as a work area for the CPU 1001, and loads various programs and data stored in the ROM 1002 and the storage unit 101. The control unit 100 executes various functional processes of the recognition device 10 by the CPU 1001 operating in accordance with the ROM 1002 and an information processing program stored in the storage unit 101 and loaded in the RAM 1003. The control unit 100 executes, for example, registration of environmental information, image recognition, and filtering of images.

[0014] The storage unit 101 is composed of a storage medium such as a hard disk drive (HDD) or a flash memory. The storage unit 101 stores an operating system for operating the recognition device 10 and other necessary software such as application programs. The storage unit 101 also stores an environment information table 1011. Note that the environment information table 1011 only needs to be stored in a storage medium or device accessible from the recognition device 10, and may also be stored in an external HDD, a management terminal or server installed in a store, or the like.

[0015] The communication I / F 102 transmits and receives data to and from devices that are communicatively connected via wired or wireless communication.

[0016] 1 is a block diagram showing an example of an imaging device 20 according to a first embodiment. The imaging device 20 is composed of, for example, a control unit 200, a storage unit 201, an imaging unit 202, and a communication I / F 203, and each unit is connected via a bus.

[0017] The control unit 200 includes a central processing unit (CPU) 2001, a read only memory (ROM) 2002, and a random access memory (RAM) 2003. The CPU 2001 controls the entire imaging device 20. The ROM 2002 stores various programs, such as a program used to drive the CPU 2001, and various data. The RAM 2003 is used as a work area for the CPU 2001, and loads various programs and data stored in the ROM 2002 and the storage unit 201. The control unit 200 executes various functional processes of the recognition device 20 by the CPU 2001 operating in accordance with the ROM 2002 and an information processing program stored in the storage unit 201 and loaded in the RAM 2003. The control unit 200 executes, for example, imaging and development processes.

[0018] The storage unit 201 is composed of a storage medium such as a hard disk drive (HDD) or a flash memory. The storage unit 201 stores an operating system for operating the imaging device 20 and other necessary software such as application programs. The storage unit 201 stores a flag for checking whether environmental information has been registered. The storage unit 201 also stores a device ID for identifying the device and generated image data.

[0019] The imaging unit 202 captures an image. The imaging unit 202 is composed of an imaging element and an optical lens, and captures an image of a preset imaging range.

[0020] The communication I / F 203 transmits and receives data to and from devices connected for communication via wired or wireless communication.

[0021] FIG. 4 is a schematic diagram showing an example of an environment information table 1011 according to the first embodiment. The environment information table 1011 stores environment information representing the environment in which each imaging device 20 is installed. In this embodiment, the environment information representing the environment in which each imaging device 20 is installed is represented by three elements, for example, luminance distribution, frequency distribution, and color temperature distribution. Therefore, the environment information table 1011 in this embodiment is composed of four areas, namely, a device ID area 50, a luminance distribution area 51, a frequency distribution area 52, and a color temperature distribution area 53. Note that, in this embodiment, the above-mentioned three elements are registered as the environment information of the environment in which the imaging device 20 is installed, but fewer elements or more elements may be registered.

[0022] The device ID area 50 stores a device ID that uniquely identifies the imaging device 20 connected to the recognition device 10. The device ID is, for example, a unique alphanumeric character or a combination thereof that is assigned in advance to each imaging device 20.

[0023] Next, the luminance distribution area 51, frequency distribution area 52, and color temperature distribution area 53, which indicate the environment, will be described. The luminance distribution area 51 is where the distribution of brightness in the installation environment is registered. The frequency distribution area 52 is where the distribution of color boundary strength in the installation environment is registered. The color temperature distribution area 53 is where the distribution of colors in the installation environment is registered.

[0024] In the luminance distribution area 51, the frequency distribution area 52, and the color temperature distribution area 53, for example, coefficients of a function indicating a surface approximation generated from pixels as an indication of the distribution of luminance and frequency of the received image are registered. For example, the coefficients are obtained by the following process. When registering the distribution of luminance in the luminance distribution area 51, the control unit 100 acquires the luminance of each pixel from the received image and calculates the average value of the luminance. After calculating the average value, the control unit 100 calculates the difference between each pixel and the average value. Based on the difference calculated between the luminance of each pixel and the average value of the luminance of the entire image, a two-dimensional matrix is ​​calculated to generate a surface approximation. After generating the surface approximation, the control unit 100 calculates a function from the surface approximation, extracts coefficients from the function, and registers the extracted coefficients as the distribution of luminance in the image.

[0025] Moreover, the device ID area 50 is registered in association with areas such as a luminance distribution area 51, a frequency distribution area 52, and a color temperature distribution area 53, which indicate the environment.

[0026] Next, a process for registering environmental information will be described. In this embodiment, a chart image is used when registering environmental information. The chart image is an image that serves as a reference when registering environmental information, and is, for example, a pure white image.

[0027] FIG. 5 is a flowchart showing an example of the environment information registration process of the imaging device 20 according to the first embodiment.

[0028] The CPU 2001 checks whether environmental information has been registered (ACT101). The CPU 2001 accesses the storage unit 201 and checks whether a flag indicating whether environmental information has been registered is valid. If the flag is valid (YES in ACT101), the CPU 2001 ends the process.

[0029] If the flag is invalid (NO in ACT101), the CPU 2001 captures an image to register the environmental information (ACT102). The CPU 2001 captures a chart image via the imaging unit 202, generates image data to be an image for environmental information registration, and stores the image data in the RAM 2003 or the storage unit 201.

[0030] After capturing the chart image, the CPU 2001 transmits the environmental information (ACT103). The CPU 2001 accesses the storage unit 201 and extracts the device ID. The CPU 2001 also accesses the RAM 2003 or the storage unit 201 and extracts an image for environmental information registration. After extracting the device ID and the image for environmental information registration, the CPU 2001 transmits the device ID and the image for environmental information registration to the recognition device 10.

[0031] Next, an environment information registration process on the recognition device side will be described with reference to a flowchart shown in FIG 6. The environment information registration process of the recognition device 10 according to the first embodiment is shown in FIG 6.

[0032] The CPU 1001 checks whether environmental information has been received (ACT201). The CPU 1001 checks whether a device ID and an image for registering environmental information have been received via the communication I / F 102. If the result of the check is that a device ID and an image for registering environmental information have not been received (NO in ACT201), a loop process is executed until a device ID and an image for registering environmental information are received.

[0033] As a result of the check, if it is confirmed that the device ID and the image for registering environmental information have been received (YES in ACT201), the CPU 1001 accesses the environmental information table 1011 and checks whether the device ID has already been registered (ACT202). If it is not confirmed that the device ID has been registered (NO in ACT202), the CPU 1001 registers the received device ID in the device ID area 50 of the environmental information table 1011.

[0034] As a result of the check, if the device ID has already been registered in the environment information table 1011 (YES in ACT202), the CPU 1001 calculates the environment information (ACT204). Taking the case of registering the luminance of an image for registering environment information as an example, the CPU 1001 acquires the luminance from each pixel of the image for registering environment information. After acquiring the luminance, the CPU 1001 calculates the average luminance value of the entire image for registering environment information based on the luminance of each pixel. After calculating the average luminance, the CPU 1001 calculates the difference between the luminance of each pixel of the image for registering environment information and the average value. The CPU 1001 calculates a two-dimensional matrix from the calculated difference, and generates a surface approximation. After generating the surface approximation, the CPU 1001 extracts coefficients from a function that represents the surface approximation.

[0035] After calculating the coefficient, the CPU 1001 registers the coefficient as environmental information (ACT 205). The CPU 1001 registers the coefficient in the environmental information table 1011 in association with the received device ID.

[0036] Next, the filtering process will be described with reference to Fig. 7. Fig. 7 is a flowchart showing an example of the filtering process according to the first embodiment. The filtering process is performed by the imaging device 20 and the recognition device 10 in cooperation with each other.

[0037] The CPU 2001 captures an image of an imaging range (ACT 301). The CPU 2001 captures an image of a preset imaging range to generate an input image to be transmitted to the recognition device 10 as image data.

[0038] The CPU 2001 transmits the input image (ACT 302). The CPU 2001 transmits the generated input image and the device ID to the recognition device 10. After transmitting the input image and the device ID, the CPU 2001 loops a process of checking whether the recognition result has been received from the authentication device 10 (NO in ACT 303) until the recognition result is received from the authentication device 10.

[0039] The CPU 1001 checks whether the input image and the device ID have been received (ACT401). If the input image and the device ID have not been received (NO in ACT401), the CPU 1001 executes a loop process until the input image and the device ID have been received.

[0040] When reception of the input image and the device ID is confirmed (YES in ACT401), the CPU 1001 acquires environmental information (ACT402). Based on the device ID received in ACT401, the CPU 1001 acquires coefficients associated with the same device ID from the environmental information table 1011.

[0041] After obtaining the coefficients, the CPU 1001 executes image filtering (ACT 403). The CPU 1001 creates a filter based on the obtained coefficients and executes filtering processing on the input image.

[0042] After completing the filtering of the image, the CPU 1001 performs image recognition on the input image (ACT 404). The CPU 1001 recognizes objects and the like that appear in the filtered input image.

[0043] After the input image is recognized, the CPU 1001 transmits the recognition result (ACT 405). The CPU 1001 identifies the imaging device 20 that transmitted the input image based on the received device ID, and transmits the recognition result. After transmitting the recognition result, the process returns to ACT 401.

[0044] If the CPU 2001 receives the recognition result (YES in ACT 303), the process returns to ACT 301.

[0045] By performing the above-described processing, the recognition device 10 performs filtering processing on images received from the connected imaging devices 20, eliminating the need to adjust each imaging device 20 or reconstruct the system.

[0046] This embodiment can be modified in various ways. For example, the filtering process in this embodiment uses one piece of environmental information, but two or more pieces of environmental information may be used. Even in this case, there is no need to adjust the imaging device 20 or reconstruct the system, and image recognition is further less affected by the installation environment, improving the accuracy of image recognition.

[0047] Furthermore, each piece of environmental information registered in the environmental information table 1011 of the recognition device 10 may be updatable.

[0048] Although several embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0049] 10…Recognition device 20...Imaging device 1011...Environment information table

Claims

1. a receiving unit that receives the first image and the second image; a control unit that calculates environmental information related to an imaging environment from the first image, creates a filter for correcting the second image, and performs control to correct the second image based on the filter; A recognition device comprising:

2. 2. The recognition device according to claim 1, wherein the control unit calculates first environmental information regarding an imaging environment from the first image and second environmental information regarding an imaging environment different from the first environmental information, creates a first filter from the first environmental information to correct the second image, creates a second filter from the second environmental information to correct the second image, and performs correction on the second image based on the first filter and the second filter.

3. The recognition device according to claim 1 , wherein the control unit performs image recognition of the second image after performing the correction based on the filter.

4. A storage unit that stores the first environmental information, The recognition device according to claim 1, wherein, when the control unit receives the first image after storing the first environmental information, the control unit calculates third environmental information regarding the imaging environment and overwrites the first environmental information stored in the memory unit with the third environmental information and stores the third environmental information.

5. A computer as a recognition device that corrects the received image and performs image recognition, a receiving means for receiving a first image and a second image; an environment information calculation means for calculating first environment information relating to an imaging environment from the first image; a filter generating means for generating a filter for correcting the second image from the first environmental information; a filtering processing means for performing correction based on the filter on the second image; A program that can be used to

Images