JP7458919B2 - Brightness prediction device and brightness prediction program - Google Patents

Description

The present invention relates to a brightness perception prediction device and a brightness perception prediction program.

In recent years, an index called brightness perception, which indicates the amount of brightness that humans perceive in a space, has been proposed. The perception of brightness is simply the amount of brightness that humans perceive, and is a different index from illuminance, which is a physical quantity that represents the brightness of light illuminating the surface of an object. Therefore, for example, even if the illuminance of lighting installed in a space is high, the perceived brightness of the space does not necessarily increase.

2. Description of the Related Art Conventionally, devices or programs have been proposed for predicting the perceived brightness of a space. For example, Patent Document 1 discloses a brightness perception prediction device that predicts the brightness perception of a target space based on the luminance of each pixel included in an image captured of the target space whose brightness perception is to be predicted. has been done. Further, Patent Document 2 discloses a brightness perception prediction device that predicts the brightness perception of a target space based on the illuminance at a point on a vertical plane in the target space.

Patent No. 3995201 International Publication No. 2012/046840 pamphlet

BACKGROUND ART Conventionally, when predicting the brightness of a target space using a device or a program, there have been cases where the predicted brightness is different from the brightness that a human actually feels from the target space. In other words, there were cases in which the brightness perception of the target space could not be appropriately predicted.

The objective of the present invention is to predict the perceived brightness of a target space with high accuracy.

The present invention includes a surface type setting unit that sets the surface type of each image area corresponding to each surface included in a brightness image representing a target space, and a weight that indicates the degree of contribution to brightness for each surface type. With reference to the associated weight information, for each of the plurality of image regions, the brightness values of pixels other than the pixels whose brightness values are equal to or higher than a predetermined value among the plurality of pixels included in the image region , and the brightness values of the pixels included in the image region are determined. The brightness feeling prediction device is characterized by comprising: a brightness feeling prediction unit that predicts the brightness feeling of the target space based on the weight of the surface type.

Preferably, the brightness perception prediction unit calculates a representative value of brightness values of a plurality of pixels included in a first image region among the plurality of image regions and a position near the first image region in the brightness image. The brightness feeling of the target space is predicted further based on a ratio of brightness values of a plurality of pixels included in the second image area to a representative value.

The present invention also provides a surface type setting section that sets the surface type of each image area corresponding to each surface included in the brightness image representing the target space, and a surface type setting section that indicates the degree of contribution to the brightness feeling for each surface type. With reference to weight information associated with weights, the brightness of the target space is determined for each of the plurality of image regions based on the luminance values of a plurality of pixels included in the image region and the weight of the surface type of the image region. a brightness perception prediction unit that predicts a brightness perception; Further, based on the ratio of the brightness values of a plurality of pixels included in a second image area located near the first image area in the brightness image to a representative value, the brightness feeling of the target space is predicted, and the brightness of the target space is predicted. The sensitivity prediction unit calculates representative values of brightness values of a plurality of pixels included in an image area whose surface type is floor among the plurality of image areas, and representative values of brightness values of a plurality of pixels included in other image areas. The present invention is characterized in that the perceived brightness of the target space is predicted without considering the ratio with the representative value.

The present invention also provides a surface type setting section that sets the surface type of each image area corresponding to each surface included in the brightness image representing the target space, and a surface type setting section that indicates the degree of contribution to the brightness feeling for each surface type. The brightness of the target space is calculated based on the luminance values of a plurality of pixels included in each of the plurality of image regions and the weight of the surface type of the image region, with reference to weight information associated with weights. a brightness perception prediction unit that predicts the brightness perception, the brightness perception prediction unit intermittently predicts the brightness perception of the target space, and predicts the brightness perception of the target space predicted immediately before. , the currently predicted brightness of the target space is corrected based on the difference between the currently predicted brightness of the target space.

The present invention also provides a computer with a surface type setting unit that sets the surface type of each image area corresponding to each surface included in a brightness image representing a target space, and a surface type setting section that sets a surface type of each image area corresponding to each surface included in a brightness image representing a target space, and a surface type setting section that sets a surface type of each image area corresponding to each surface included in a brightness image representing a target space, and a surface type setting section that sets a surface type of each image area corresponding to each surface included in a brightness image representing a target space. For each of the plurality of image regions, the brightness values of pixels other than those whose brightness values are equal to or higher than a predetermined value, among the plurality of pixels included in the image regions , A brightness perception prediction program is characterized in that the program functions as a brightness perception prediction unit that predicts the brightness perception of the target space based on the weight of the surface type of the image area .

The present invention makes it possible to predict the perceived brightness of a target space with high accuracy.

1 is a schematic diagram illustrating the configuration of a brightness perception prediction device according to an embodiment of the present invention. FIG. 13 is a diagram illustrating an example of a luminance image. FIG. 3 is a conceptual diagram showing an example of weight information. FIG. 3 is a diagram showing image regions defined in a luminance image.

FIG. 1 is a schematic diagram of the configuration of a brightness perception prediction device 10 according to this embodiment. The brightness perception prediction device 10 is, for example, a server computer, a personal computer, or a tablet terminal, but may be any device as long as it can perform the functions described below. Further, although the brightness perception prediction device 10 according to the present embodiment exhibits each function described below with one device, each function of the brightness perception prediction device 10 may be performed by a plurality of devices.

The communication unit 12 includes, for example, a network adapter. The communication unit 12 exhibits a function of communicating with other devices (for example, a camera, an image processing device that processes images captured by the camera, etc.) via a communication line such as a LAN or the Internet. For example, the communication unit 12 receives a luminance image 20, which will be described later, from another device. Furthermore, the communication unit 12 may transmit the prediction result by the brightness perception prediction unit 28, which will be described later, to another device.

The display unit 14 includes, for example, a liquid crystal display or an organic EL display. The display unit 14 displays various screens. For example, the display unit 14 can display the brightness image 20 and the prediction result by the brightness perception prediction unit 28.

The input unit 16 includes, for example, a mouse, a keyboard, a touch panel, a microphone, and the like. The input unit 16 is used to input instructions from the user to the brightness perception prediction device 10.

The storage unit 18 is configured to include, for example, a hard disk, SSD, ROM, or RAM. The storage unit 18 stores a brightness perception prediction program for operating each part of the brightness perception prediction device 10. Further, as shown in FIG. 1, the storage unit 18 stores a luminance image 20 and weight information 22.

The luminance image 20 is an image representing a target space, which is a space for which the brightness perception prediction device 10 predicts the brightness perception. In this specification, the target space is a space that includes multiple types of surfaces (e.g., wall surfaces, ceiling surfaces, floor surfaces, etc.). For example, the target space is an indoor space.

An example of a brightness image 20 is shown in FIG. The brightness image 20 is a two-dimensional arrangement of a plurality of pixels. Each pixel included in the brightness image 20 only needs to have at least a brightness value as a pixel value. The brightness image 20 may be a camera image obtained by photographing the target space with a camera. Furthermore, while the camera image has a relatively narrow range of brightness values for each pixel (for example, 0 to 255), the brightness image 20 is a brightness image that has been processed to expand the brightness range of each pixel in the camera image. Preferably, it is an extended image. Note that the brightness image 20 may be an image formed by a computer program, such as a CG image, instead of an image based on a camera image. The brightness perception prediction device 10 receives the brightness image 20 from another device through the communication unit 12 described above and stores it in the storage unit 18 . Alternatively, the brightness perception prediction device 10 may form the brightness image 20.

Returning to FIG. 1, the weight information 22 is information in which a weight indicating the contribution to the sense of brightness is associated with each surface type that can be defined in the luminance image 20. As described above, the target space contains multiple types of surfaces, and the inventors of the present invention have found, based on survey results such as questionnaires, that the influence that the luminance value of each pixel contained in a surface has on the sense of brightness of the target space, i.e., the contribution to the sense of brightness, differs depending on the surface type contained in the target space. Therefore, weights indicating the contribution to the sense of brightness of each surface type based on the survey results are stored in advance in the storage unit 18 as weight information 22.

An example of the weight information 22 is shown in FIG. FIG. 3 shows a ceiling surface, a front wall surface, a side wall surface, and a floor surface as surface types. The ceiling surface is a surface corresponding to the ceiling of the target space, the front wall surface is a surface corresponding to a wall extending in the horizontal direction of the brightness image 20 within the field of view of the brightness image 20, and the side wall surface is a surface corresponding to the wall that extends in the left-right direction of the brightness image 20 within the field of view of the brightness image 20. Within the field of view of the image 20, this is a surface corresponding to a wall extending in the depth direction, and the floor surface is a surface corresponding to the floor of the target space. Of course, the surface types are not limited to these, and may include other types of surfaces (for example, window surfaces).

The inventors of the present invention have found that, of the four surface types shown in FIG. 3, the contribution of the surface to the sense of brightness in the target space is, in descending order, the front wall surface, the ceiling surface, the side wall surface, and the floor surface. Therefore, of the four surface types, the weight of the front wall surface is the greatest, the weight of the ceiling surface is the next greatest, the weight of the side wall surface is the next greatest, and the weight of the floor surface is the least. The actual weight values are examples, and may be changed as appropriate depending on the formula for calculating the sense of brightness, etc.

Returning to FIG. 1, the control unit 24 includes various control devices such as a CPU, GPU, ASIC, and FPGA. The control unit 24 controls each part of the brightness perception prediction device 10 based on the brightness perception prediction program stored in the storage unit 18 . In particular, as shown in FIG. 1, the control unit 24 functions as a surface type setting unit 26 and a brightness prediction unit 28.

The surface type setting unit 26 first defines each image area corresponding to each surface included in the brightness image 20. In this embodiment, the surface type setting unit 26 defines a plurality of image regions by performing image analysis on the brightness image 20. Specifically, the surface type setting unit 26 performs edge detection processing on the brightness image 20, and converts the area surrounded by the detected edge, or the area surrounded by the detected edge and the side of the brightness image 20 into one. defined as one image area. Note that at this time, an area smaller than a predetermined area (number of pixels) may be ignored without being defined as an image area. Thereby, objects other than surfaces included in the brightness image 20 (for example, furniture, pillars, etc.) can be excluded. Further, the surface type setting unit 26 may define a plurality of image regions within the brightness image 20 according to instructions from the user without performing image processing on the brightness image 20.

Next, the surface type setting unit 26 sets the surface types of the plurality of defined image areas. In this embodiment, the surface type setting unit 26 sets the surface type based on the shape of the defined image area and the position within the brightness image 20. For example, the surface type setting section 26 is a trapezoidal image area, the upper base and the lower base are parallel, the upper base is longer than the lower base, and the image area is located at the upper side of the brightness image 20. Set the surface type to ceiling surface. Further, the surface type setting unit 26 sets the surface type of the image area, which is a rectangular image area and at least one of the left and right sides is vertical (extends in the vertical direction in the brightness image 20), to be a side wall surface. Set. Further, the surface type setting unit 26 may specify the surface type of each image region according to an instruction from the user, regardless of the shape of the defined image region or the position within the brightness image 20.

FIG. 4 shows a plurality of image areas 40 defined in the brightness image 20 shown in FIG. 2 and the surface type of each image area 40. In FIG. 4, in the brightness image 20, a ceiling surface 40a, a front wall surface 40b, a side wall surface 40c, and a floor surface 40d are set.

The brightness prediction unit 28 predicts the brightness of the target space represented by the brightness image 20. It is known that the brightness perceived by humans in a target space is proportional to the logarithm of the brightness of each part of the target space. Therefore, the brightness feeling prediction unit 28 basically calculates the brightness feeling of the target space as a numerical value based on the geometric mean brightness value which is the geometric mean value of the brightness values of the plurality of pixels constituting the brightness image 20. Predict. Since various formulas have been proposed in the past for calculating the predicted value of brightness, a detailed explanation will be omitted here. Note that it is known that the light source (for example, a light bulb) included in the brightness image 20 has a fairly small influence on the brightness perception, so the brightness perception prediction unit 28 estimates that the pixel corresponding to the light source has a brightness Not used to predict feelings. Specifically, the brightness prediction unit 28 regards a pixel whose brightness value is a predetermined value or more as a pixel corresponding to a light source, and calculates the brightness based on the geometric mean value of the brightness values of other pixels excluding the pixel. Predict feelings.

In particular, in the present embodiment, the brightness prediction unit 28 refers to the weight information 22 and calculates the representative value (for example, the geometric mean) of the brightness values of the plurality of pixels included in each image area 40 set by the surface type setting unit 26. The perceived brightness of the target space represented by the brightness image 20 is predicted based on the brightness value (hereinafter referred to as "representative brightness value") and the weight of the surface type of each image region 40. Specifically, the brightness feeling prediction unit 28 changes the degree of influence of the brightness representative value of each image region 40 on the brightness feeling of the target space according to the weight of each surface type indicated by the weight information 22. , predict the brightness of the target area. In an extreme example, if there is an image region 40 with a weight of 0, the brightness prediction unit 28 ignores the representative brightness value of the image region 40 and predicts the brightness of the target space.

As an example, the brightness perception prediction unit 28 first calculates a luminance representative value of each image region 40, such as a geometric mean luminance value, for each image region 40. It then predicts the brightness perception of the target space by placing more importance on the luminance representative value of an image region 40 with a larger weight, or in other words, by placing less importance on the luminance representative value of an image region 40 with a smaller weight.

For example, assuming that the weight information 22 is as shown in FIG. ), and when the representative brightness value of the front wall surface 40b is small, it acts largely in the direction where the brightness of the target space decreases (it feels dark). On the other hand, since the weight of the floor surface 40d is small, even if the representative luminance value of the floor surface 40d is large, it does not have much effect on increasing the brightness of the target space, and the representative luminance value of the floor surface 40d Even when the value is small, it does not have much effect on reducing the perceived brightness of the target space.

Furthermore, it is preferable that the brightness perception prediction unit 28 predicts the brightness perception of the target space further based on the ratio between the luminance representative value of the first image region 40 in the luminance image 20 and the luminance representative value of the second image region 40 located near the first image region 40 in the luminance image 20. Here, the second image region 40 located near the first image region 40 is typically an image region 40 adjacent to the first image region 40, but the second image region 40 does not necessarily have to be adjacent to the first image region 40 as long as it is located near the first image region 40.

The inventor of the present invention believes that in the brightness image 20, the larger the ratio of the brightness representative values of two nearby image areas 40, the darker the target space is perceived by humans, that is, the lower the sense of brightness is. I discovered that. Note that a large ratio between the representative brightness values of the two image areas 40 means that the difference between the representative brightness values of both the image areas 40 is large. Referring to FIG. 2, for example, if the ratio of the representative brightness values between the central ceiling surface and the side walls adjacent to the central ceiling surface is large, the ratio of the typical brightness values between the ceiling surface and the side walls adjacent to the central ceiling surface is large. Humans feel that the target space is darker than when the distance is small.

Therefore, the brightness perception prediction unit 28 determines whether each image area 40 (each first image area 40) included in the brightness image 20 is different from each other in the vicinity of each of the other image areas 40 (second image area 40). Calculate the ratio of representative brightness values between. Then, the brightness of the target space is predicted by considering the sum or average value of the plurality of calculated ratios. Specifically, the larger the sum or average value of multiple calculated ratios, the lower the brightness of the target space, and the smaller the sum or average value of multiple calculated ratios, the lower the brightness of the target space. Enlarge.

Generally, the representative luminance value of the floor surface 40d tends to be quite low. Therefore, even if a person feels that the target space is relatively bright, the image area 40 corresponding to the floor surface 40d and other image areas 40 in the vicinity (for example, the front wall surface 40b and the side wall surface 40c) The ratio of representative brightness values between the image area 40) corresponding to the image area 40) tends to increase. Therefore, when calculating the ratio of representative brightness values between the two image regions 40, the brightness feeling prediction section 28 preferably does not consider the image region 40 corresponding to the floor surface 40d set by the surface type setting section 26. It is. In other words, the brightness perception prediction unit 28 calculates the brightness of the target space without considering the ratio between the brightness representative value of the image area 40 corresponding to the floor surface 40d and the brightness representative values of other image areas 40. It is preferable to predict the feeling.

Further, the brightness perception prediction unit 28 predicts the brightness perception of the same target space intermittently, in other words, at a plurality of different times. For example, the brightness prediction unit 28 predicts the brightness of the target space periodically, such as every 30 minutes or every 5 minutes. This is because the brightness of the target space may change depending on the time of day. For example, in the daytime, sunlight serves as a light source, so the position of the sun changes depending on the time of day, and as a result, the way the light hits the target space changes, causing the perceived brightness to fluctuate. Furthermore, at night, the main light source is a lighting fixture, so the brightness in the same target space may be completely different from that in the daytime.

The brightness prediction unit 28 intermittently predicts the brightness of the target space based on the plurality of brightness images 20 that are intermittently acquired. For example, a camera that images a certain target space is fixedly installed, and the camera intermittently photographs the target space to obtain a plurality of camera images. The plurality of camera images may be used as the plurality of brightness images 20, or the plurality of brightness expanded images subjected to processing for expanding the brightness range of each camera image may be used as the plurality of brightness images 20.

When the space changes from a bright space to a dark space, the human eye is not used to the darkness and has difficulty seeing the surroundings at first, but as time passes, the eyes gradually become accustomed to the surroundings and the surroundings become more visible. It becomes easier to see (dark adaptation). Also, when the space changes from a dark space to a bright space, the human eye is not accustomed to the brightness and has difficulty seeing the surroundings at first, but as time passes, the eyes gradually become accustomed to the brightness. It becomes easier to see your surroundings (light adaptation).

When predicting the sense of brightness of the target space intermittently, the sense of brightness prediction unit 28 may predict the sense of brightness of the target space taking into account the above-mentioned dark adaptation or light adaptation. Specifically, the sense of brightness prediction unit 28 may correct the sense of brightness of the target space predicted this time based on the difference between the sense of brightness of the target space predicted immediately before and the sense of brightness of the target space predicted this time.

Specifically, the brightness perception prediction unit 28 calculates that the larger the difference obtained by subtracting the currently predicted brightness of the target space from the brightness of the target space predicted immediately before, the brightness of the currently predicted target space. Correct to reduce the feeling. In other words, the brightness perception prediction unit 28 determines that the smaller the difference obtained by subtracting the currently predicted brightness of the target space from the brightness of the target space predicted immediately before, the brightness perception of the target space predicted this time. Correct to increase the brightness.

For example, suppose that the sense of brightness is expressed by a numerical value from 0 to 20, and the brightness of the currently predicted target space is 5 (that is, it is predicted to be relatively dark). At this time, the brightness perception prediction unit 28 predicts the brightness of the target space as compared to the case where the brightness perception of the target space predicted immediately before is 6, that is, the difference in brightness perception is 1. When the perceived brightness of the target space is 12, that is, when the difference in perceived brightness is 7, the correction is made so that the larger the difference, the smaller the predicted value of the current perceived brightness. do.

Also, assume that the brightness of the currently predicted target space is 15 (i.e., predicted to be relatively bright). At this time, the brightness perception prediction unit 28 performs a correction so that the current predicted value of brightness perception is larger when the previously predicted brightness of the target space is 9, i.e., the difference in brightness perception is -6, i.e., the smaller the difference, compared to when the previously predicted brightness of the target space is 14, i.e., the difference in brightness perception is -1.

The outline of the configuration of the brightness perception prediction device 10 according to this embodiment is as described above. According to the brightness perception prediction device 10, weights are set according to the multiple surface types defined in the luminance image 20, and the brightness perception of the target space is predicted based on the luminance image 20 while taking into account the weights. This makes it possible to predict the brightness perception with high accuracy, i.e., closer to the actual human perception, compared to a case where the brightness perception is predicted without taking into account at least the surface types included in the luminance image 20.

Although the embodiments according to the present invention have been described above, the present invention is not limited to the above embodiments, and various changes can be made without departing from the spirit of the present invention.

10 brightness prediction device, 12 communication unit, 14 display unit, 16 input unit, 18 storage unit, 20 brightness image, 22 weight information, 24 control unit, 26 surface type setting unit, 28 brightness prediction unit, 40 image Area, 40a ceiling surface, 40b front wall surface, 40c side wall surface, 40d floor surface.

Claims (5)

a surface type setting unit that sets a surface type of each image region corresponding to each surface included in a luminance image representing a target space;
a brightness perception prediction unit that refers to weight information in which a weight indicating a contribution of each surface type to a brightness perception is associated with each surface type , and predicts a brightness perception of the target space based on the brightness values of pixels other than pixels having a brightness value equal to or greater than a predetermined value among the pixels included in each of the plurality of image regions, and the weights of the surface types of the image regions;
A brightness perception prediction device comprising: The brightness perception prediction unit predicts the brightness perception of the target space based on a ratio between a representative value of luminance values of a plurality of pixels included in a first image region among the plurality of image regions and a representative value of luminance values of a plurality of pixels included in a second image region located near the first image region in the luminance image.
2. The brightness perception prediction device according to claim 1, a surface type setting unit that sets the surface type of each image area corresponding to each surface included in the brightness image representing the target space;
Referring to weight information in which weights indicating the degree of contribution to brightness are associated with each surface type, the luminance values of a plurality of pixels included in the image region and the image for each of the plurality of image regions are determined. a brightness feeling prediction unit that predicts the brightness feeling of the target space based on the weight of the surface type of the area;
Equipped with
The brightness prediction unit is configured to calculate representative values of brightness values of a plurality of pixels included in a first image region among the plurality of image regions, and a second image region located in the vicinity of the first image region in the brightness image. Predicting the brightness of the target space further based on the ratio of the brightness values of a plurality of pixels included in the image area to a representative value,
The brightness perception prediction unit calculates representative values of brightness values of a plurality of pixels included in an image area whose surface type is a floor surface among the plurality of image areas, and brightness values of a plurality of pixels included in other image areas. predicting the brightness of the target space without considering the ratio of the value to the representative value;
A brightness perception prediction device characterized by: a surface type setting unit that sets the surface type of each image area corresponding to each surface included in the brightness image representing the target space;
Referring to weight information in which weights indicating the degree of contribution to brightness are associated with each surface type, the luminance values of a plurality of pixels included in the image region and the image for each of the plurality of image regions are determined. a brightness feeling prediction unit that predicts the brightness feeling of the target space based on the weight of the surface type of the area;
Equipped with
The brightness feeling prediction unit is
Intermittently predicting the brightness of the target space,
correcting the currently predicted brightness of the target space based on the difference between the brightness of the target space predicted immediately before and the currently predicted brightness of the target space;
A brightness perception prediction device characterized by: computer,
a surface type setting unit that sets the surface type of each image area corresponding to each surface included in the brightness image representing the target space;
By referring to weight information in which a weight indicating the degree of contribution to brightness is associated with each surface type, a predetermined brightness value of a plurality of pixels included in the plurality of image regions is determined for each of the plurality of image regions. a brightness feeling prediction unit that predicts the brightness feeling of the target space based on the brightness value of pixels other than the pixel that is equal to or greater than the value and the weight of the surface type of the image area;
A brightness perception prediction program that is characterized by functioning as a.