JP2018142828A - Deposit detector and deposit detection method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a deposit detector and a deposit detection method which enable the detection of an opaque deposit.SOLUTION: A deposit detector according to an embodiment comprises a calculating unit and an extraction unit. The calculating unit calculates hue values of pixels included in a taken image shot by an imaging device. The extraction unit extracts a deposit region indicative of a shooting region of a deposit depositing on a lens of the imaging device based on a hue distribution which is a distribution of hue values calculated by the calculating unit.SELECTED DRAWING: Figure 1B

Description

  The present invention relates to a deposit detection apparatus and a deposit detection method.

  2. Description of the Related Art Conventionally, for example, there is an attached matter detection device that detects attached matter attached to a lens of an in-vehicle camera attached to a vehicle. Such an adhering matter detection device extracts the luminance edge of the adhering matter from the captured image, and detects a water droplet adhering to the lens as an adhering matter (see, for example, Patent Document 1).

JP 2010-14494 A

  However, in the conventional attached matter detection method, water droplets can be detected as the attached matter, but for example, an opaque attached matter that does not transmit light such as dust cannot be detected.

  This invention is made in view of the above, Comprising: It aims at providing the deposit | attachment detection apparatus and the deposit | attachment detection method which can detect an opaque deposit | attachment.

  The present invention includes a calculation unit and an extraction unit in an attached matter detection apparatus. The calculation unit calculates the hue value of each pixel included in the captured image captured by the imaging device. The extraction unit extracts an attachment region indicating an imaging region of the attachment attached to the lens of the captured image based on the hue distribution that is the distribution of the hue values calculated by the calculation unit.

  According to the present invention, an opaque deposit can be detected.

FIG. 1A is a diagram illustrating a mounting example of an attached matter detection apparatus. FIG. 1B is a diagram showing an outline of the deposit detection method. FIG. 2 is a block diagram of the adhering matter detection apparatus. FIG. 3A is a diagram (part 1) illustrating a specific example of processing by the generation unit. FIG. 3B is a second diagram illustrating a specific example of the process performed by the generation unit. FIG. 3C is a third diagram illustrating a specific example of the process performed by the generation unit. FIG. 4A is a diagram (part 1) illustrating a specific example of processing by the extraction unit. FIG. 4B is a second diagram illustrating a specific example of processing performed by the extraction unit. FIG. 4C is a third diagram illustrating a specific example of the process performed by the extraction unit. FIG. 5 is a flowchart illustrating a processing procedure executed by the attached matter detection apparatus. FIG. 6 is a diagram illustrating processing contents of the attached matter detection device according to the modification.

  Hereinafter, an attached matter detection apparatus and an attached matter detection method according to an embodiment will be described in detail with reference to the accompanying drawings. In addition, this invention is not limited by this embodiment.

  First, the outline | summary of the deposit | attachment detection apparatus and deposit | attachment detection method which concern on embodiment is demonstrated using FIG. 1A and FIG. 1B. FIG. 1A is a diagram illustrating a mounting example of an attached matter detection apparatus. FIG. 1B is a diagram showing an outline of the deposit detection method.

  As shown in FIG. 1A, the deposit detection device 1 is mounted on a vehicle C. Further, the attached matter detection apparatus 1 detects the attached matter F attached to the lens 10a (not shown) of the camera 10 from the captured image L captured by the camera 10 which is an imaging device mounted on the vehicle C.

  In the example shown in the figure, a case where the vehicle C includes four cameras 10 that capture images in different directions is illustrated. Here, since the camera 10 is disposed outside the vehicle C, there is a possibility that an adhering substance F such as water droplets or dust may adhere to the lens 10a of the camera 10.

  However, in the conventional attached matter detection apparatus, only the water droplet is detected as the attached matter F, and detection of the attached matter F other than the water droplet is not considered.

  Therefore, in the attached matter detection method according to the embodiment, the opaque attached matter F attached to the lens 10a of the camera 10 is detected. That is, in the adhering matter detection method according to the embodiment, the opaque adhering matter F adhering to the lens 10a of the camera 10 is detected by paying attention to the hue distribution that is the distribution of the hue value of each pixel included in the captured image L. To do.

  Specifically, as illustrated in FIG. 1B, first, in the attached matter detection method according to the embodiment, the hue value of each pixel included in the captured image L is calculated (step S1). Here, in the deposit region M where the opaque deposit F adheres in the captured image L, light is blocked by the deposit F, and therefore in the deposit region M due to noise or the like in the image sensor of the camera 10. Hue values of specific colors appear randomly.

  In the attached matter detection method according to the embodiment, paying attention to such points, the attached matter region M is extracted based on the hue distribution (step S2). Specifically, in the attached matter detection method according to the embodiment, a region where a specific hue value appears at random is extracted as the attached matter region M.

  That is, in the attached matter detection method according to the embodiment, the opaque attached matter F attached to the lens 10a of the camera 10 is detected using the feature that the hue distribution of the attached region M is disturbed.

  Thereby, in the deposit | attachment detection method which concerns on embodiment, the opaque deposit | attachment F can be detected.

  By the way, the attached matter detection method according to the embodiment generates a binarized hue image Lh obtained by binarizing each pixel based on the hue value, and extracts the attached matter region M from the binarized hue image Lh. it can. Details of this point will be described later with reference to FIG. 3A and the like.

  In the attached matter detection method according to the embodiment, the attached matter region M can be extracted based on the brightness value and the saturation value of each pixel in addition to the hue value. Details of this point will be described later with reference to FIGS. 3B and 3C.

  In the attached matter detection method according to the embodiment, the attached matter region M can also be determined based on a plurality of time-series captured images L. Details of this point will be described later with reference to FIG. 4C.

  Next, the structure of the deposit | attachment detection apparatus 1 which concerns on embodiment is demonstrated using FIG. FIG. 2 is a block diagram of the deposit detection apparatus 1. In the figure, the camera 10 and the deposit removing device 11 are shown together.

  The camera 10 is an in-vehicle camera including an image sensor (imaging device) such as a CMOS (Complementary Metal Oxide Semiconductor), for example, and images the surroundings of the vehicle C as described above. A captured image L captured by the camera 10 is output to the attached matter detection apparatus 1. The camera 10 may use another image sensor such as a CCD (Charge Coupled Device) instead of the CMOS. The lens 10a of the camera 10 is, for example, a wide-angle lens such as a fisheye lens, but may be other lenses.

  The attached matter removing device 11 performs an operation for removing the attached matter attached to the lens 10 a of the camera 10 based on the removal operation instruction from the attached matter detecting device 1. The deposit removing device 11 can remove deposits adhering to the lens 10a by, for example, jetting compressed air or washer liquid toward the lens 10a of the camera 10 or wiping the lens 10a with a wiper. .

  The attached matter detection apparatus 1 includes a control unit 2 and a storage unit 3. The control unit 2 includes a calculation unit 21, a generation unit 22, and an extraction unit 23. The control unit 2 includes, for example, a computer having a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), a hard disk drive (HDD), an input / output port, and various circuits.

  The CPU of the computer functions as the calculation unit 21, the generation unit 22, and the extraction unit 23 of the control unit 2, for example, by reading and executing a program stored in the ROM.

  In addition, at least one or all of the calculation unit 21, the generation unit 22, and the extraction unit 23 of the control unit 2 may be configured by hardware such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). it can.

  The storage unit 3 corresponds to, for example, a RAM or an HDD. The RAM and HDD can store specific color information 31, threshold information 32, point information 33, and various program information.

  In addition, the deposit | attachment detection apparatus 1 is good also as acquiring the above-mentioned program and various information via the other computer and portable recording medium which were connected by the wired or wireless network.

  The calculation unit 21 of the control unit 2 calculates the hue value of each pixel included in the captured image L input from the camera 10. Specifically, the calculation unit 21 performs the HSV conversion on the captured image L to calculate the hue value (so-called H value) of each pixel included in the captured image L. For example, the hue value is represented by 180 numbers from 0 to 179.

  The calculation unit 21 can also calculate a saturation value (so-called S value) and a lightness value (so-called V value) in addition to the hue value by performing HSV conversion on the captured image L. For example, the saturation value and the lightness value are represented by numbers from 0 to 255, respectively.

  The calculation unit 21 generates image data in which the calculated hue value, saturation value, and brightness value of each pixel are associated with each corresponding pixel, and outputs the image data to the generation unit 22.

  In addition, although the case where the calculation unit 21 divides the hue value into 180 pieces has been described here, the present invention is not limited to this, and may be divided into 180 pieces or more, or may be divided into less than 180 pieces. You may make it do. Further, the calculation unit 21 may calculate a hue value or the like by using another conversion method such as HLS conversion instead of the HSV conversion.

  The generation unit 22 generates a binarized hue image Lh in which each pixel is binarized depending on whether the hue value calculated by the calculation unit 21 is a specific color unique to the deposit region M or other than the specific color. Generate and output to the extraction unit 23.

  The generation unit 22 can also generate a binary lightness image Lv and a binary saturation image Ls corresponding to the lightness value and the saturation value calculated by the calculation unit 21, respectively. The binarized lightness image Lv and the binarized chroma image Ls are output to the extraction unit 23. In the following description, the binarized hue image Lh, the binarized lightness image Lv, and the binarized chroma image Ls generated by the generating unit 22 may be collectively referred to simply as “binarized image”. .

  Here, the specific example of the process by the production | generation part 22 is demonstrated using FIG. 3A-FIG. 3C. 3A to 3C are diagrams illustrating specific examples of processing by the generation unit 22.

  First, the case where the generation unit 22 generates the binarized hue image Lh based on the hue value will be described with reference to FIG. 3A. In FIG. 3A, a part of the image data is cut out and schematically shown for easy explanation.

  As shown in FIG. 3A, each pixel of the image data is associated with a numerical value of the hue value calculated by the calculation unit 21. Then, the generation unit 22 sets a pixel that is a hue value of a specific color indicating the feature of the adhered region M to “1” in the image data, and sets a pixel that is a hue value other than the specific color to “0”. Thus, the binarized hue image Lh is generated. In the figure, in the binarized hue image Lh, pixels corresponding to the specific color are shown in black, and pixels not corresponding to the specific color are shown in white.

  Here, the information regarding the hue value of the specific color is created in advance using experiments, machine learning, or the like, and is stored as the specific color information 31 in the storage unit 3 (see FIG. 2). The specific color information 31 is, for example, information obtained by extracting the top 20 colors from hue values that have a high probability of appearing in the deposit region M.

  As described above, the generation unit 22 generates the binarized hue image Lh based on the hue value, so that pixels having a plurality of hue values that are specific colors can be made equivalent. In other words, the binarized hue image Lh is a map in which the hue distribution of each pixel included in the captured image L is simplified. Thereby, the process by the latter extraction part 23 can be made easy.

  Next, a case where the generation unit 22 generates a binarized brightness image Lv from image data based on the brightness value will be described with reference to FIG. 3B. As illustrated in FIG. 3B, the generation unit 22 binarizes each pixel by comparing the brightness value of each pixel with a threshold value Tha, and generates a binarized brightness image Lv.

  Specifically, the generation unit 22 sets a pixel having a lightness value lower than the threshold value Tha to “1” and sets a pixel having a lightness value equal to or greater than the threshold value Tha to “0”, whereby the binarized lightness image Lv. Can be generated. In other words, the generation unit 22 can send only pixels with low brightness to the processing by the subsequent extraction unit 23.

  This is because the opaque deposit F does not transmit light, and therefore the brightness value of the deposit region M where the deposit F is adhered is darker than the region where the deposit F is not adhered. That is, the production | generation part 22 can improve the extraction precision of the deposit | attachment area | region M by the extraction part 23 mentioned later by producing | generating the binarized lightness image Lv based on a lightness value.

  When the brightness value is 0 to 255, the threshold value Tha is 30. Further, the generation unit 22 can change the threshold value Tha according to the illuminance around the camera 10 and the like.

  Next, a process when the generation unit 22 generates the binarized saturation image Ls from the image data based on the saturation value will be described with reference to FIG. 3C. As illustrated in FIG. 3C, the generation unit 22 determines whether or not the saturation value of each pixel is within a predetermined range Rs, thereby generating a binary saturation image Ls.

  Specifically, when the saturation value of the pixel is within the range Rs, the generation unit 22 sets the pixel to “1” and sets the pixel whose saturation value is outside the range Rs to “0”. A binary saturation image Ls is generated.

  For example, when a wide-angle lens is used as the lens 10 a of the camera 10, the edge of the lens 10 a of the camera 10 is reflected in the captured image L. This is because the edge of the lens 10a has a feature in which the hue distribution in the captured image L is the same as that of the deposit region M, while the saturation value deviates from the range Rs.

  In other words, the generation unit 22 generates a binary saturation image Ls based on the saturation value, so that the pixel corresponding to the edge of the lens 10a of the camera 10 captured in the captured image L is extracted by the subsequent extraction unit 23. It can be excluded from the extraction process. Thereby, the detection precision of the deposit | attachment F by the extraction part 23 can be improved.

  When the saturation value is 0 to 255, the range Rs is a range from 1 to 254, but the generation unit 22 can be changed as appropriate according to the illuminance around the camera 10 and the like. Note that the information on the threshold value Tha and the range Rs shown in FIG. 3B and FIG. 3C is stored as threshold information 32 in the storage unit 3.

  Returning to the description of FIG. 2, the extraction unit 23 of the control unit 2 will be described. Based on the hue distribution that is the hue value distribution calculated by the calculation unit 21, the extraction unit 23 extracts an attachment region M that indicates an imaging region of the attachment F attached to the lens 10 a of the camera 10.

  Here, the specific example of the process by the extraction part 23 is demonstrated using FIG. 4A-FIG. 4C. 4A to 4C are diagrams illustrating specific examples of processing performed by the extraction unit 23.

  As illustrated in FIG. 4A, the extraction unit 23 divides the captured image L into a plurality of regions, and extracts an attachment region M for each of the divided regions. As shown in the figure, for example, the region divides the peripheral portion of the captured image L larger than the peripheral portion.

  Specifically, the peripheral region Bl is divided into, for example, four times the size of the region Bs other than the peripheral region. Further, as will be described later with reference to FIG. 4B and FIG. 4C, the extraction unit 23 extracts the region as an attachment region M when the binary value ratio in the divided region satisfies a predetermined condition. That is, the extraction unit 23 calculates the ratio of candidate pixels that are candidates for the attachment region M for each region, and extracts the attachment region M based on the ratio.

  At this time, the region B1 having a large region size has a stricter condition of being extracted as the deposit region M than the region Bs having a smaller size than the region Bl. For example, when the same number of candidate pixels exist in the region Bl and the region Bs, the ratio of candidate pixels is smaller in the region Bl than in the region Bs. That is, the region Bl is less likely to be extracted as the adhered region M than the region Bs.

  This is because the peripheral region Bl of the captured image L has a lower priority for detecting the deposit F than the region Bs other than the peripheral portion. That is, even if the deposit F adheres to the peripheral area B1, for example, it does not hinder the driver of the vehicle C.

  Thus, the extraction part 23 can suppress the erroneous detection of the deposit | attachment F in area | region Bl by enlarging size of area | region Bl of the peripheral part compared with other area | region Bs. Note that the extraction unit 23 may exclude the peripheral portion of the captured image L from the detection target of the deposit F.

  Although the case where the region is divided into rectangles has been described here, the extraction unit 23 may divide the region into other shapes such as a circle and a polygon. Further, the extraction unit 23 may divide all the regions equally, and may perform weighting so that the extraction conditions for the attachment region M become more severe as the region becomes closer to the edge of the captured image L.

  By the way, in the conventional deposit | attachment detection apparatus, as mentioned above, the deposit | attachment F makes the water drop a detection target. For this reason, in the conventional deposit | attachment detection apparatus, the shape of a water droplet can be limited to a substantially circular shape. For this reason, in the conventional deposit | attachment detection apparatus, a water droplet can be detected by the matching process with the circular template prepared previously.

  On the other hand, in the adhering matter detection apparatus 1 according to the embodiment, the opaque adhering matter F is detected, and the irregular adhering matter F such as dust and mud is detected. For this reason, when preparing a template as described above, it is necessary to prepare an enormous number of templates. In other words, it is difficult to detect the irregular deposit F by the matching process using the template as in the prior art.

  For this reason, in the deposit | attachment detection apparatus 1 which concerns on embodiment, it becomes possible to detect the irregular-shaped deposit | attachment F by detecting the deposit | attachment F for every area | region.

  Next, a specific example of processing by the extraction unit 23 will be described with reference to FIGS. 4B and 4C. As illustrated in FIG. 4B, first, the extraction unit 23 extracts an overlapping area from the three types of binarized images input from the generation unit 22.

  Specifically, the extraction unit 23 extracts pixels that are “1” in all of the binarized hue image Lh, the binarized lightness image Lv, and the binarized chroma image Ls, and generates an adhesion determination image Ld. To do.

  In other words, the extraction unit 23 extracts an overlapping region that is a logical product of the binarized hue image Lh, the binarized lightness image Lv, and the binarized chroma image Ls, and generates an adhesion determination image Ld. That is, the extraction unit 23 extracts pixels whose brightness value is less than the threshold value Tha, saturation value is in the range Rs, and hue value is a specific color.

  Then, the extraction unit 23 calculates the ratio of pixels that are overlapping areas for each of the areas Bl and Bs illustrated in FIG. 4A in the adhesion determination image Ld, and the ratio of the pixels is equal to or higher than a threshold (for example, 80% or higher). ) Can be extracted as the deposit region M.

  Here, the reason why the ratio of the pixels to be the overlapping region in each region is that when the attached matter F is attached, the attached matter F appears in the captured image L over a plurality of pixels. In other words, it is difficult to think that the deposit F appears in only one pixel.

  That is, the extraction unit 23 extracts the adhered region M based on the ratio of pixels that are overlapping regions in each region. Pixels accidentally extracted as regions can be excluded.

  Here, the case has been described where the extracting unit 23 extracts the overlapping region by taking the logical product of the binarized hue image Lh, the binarized lightness image Lv, and the binarized chroma image Ls. 23 may extract the overlap region by taking the logical product of the binarized hue image Lh and the binarized lightness image Lv or the binarized chroma image Ls. That is, the extraction unit 23 may extract pixels whose brightness value is less than the threshold value Tha and whose hue value is a specific color as an overlapping region, or the saturation value is within the range Rs. In addition, pixels whose hue value is a specific color may be extracted as an overlapping region.

  By the way, the extraction unit 23 can determine the attachment region M based on the captured images L of a plurality of frames in time series. Specifically, as illustrated in FIG. 4C, the extraction unit 23 extracts an attachment region M based on a plurality of time-series attachment determination images Ld.

  For example, as illustrated in FIG. 4C, when the extraction unit 23 extracts the above-described attachment region M, the extraction unit 23 adds points to the region that becomes the attachment region M for each attachment determination image Ld. Then, the extraction unit 23 causes the storage unit 3 to store the cumulative value of the points in each region as the point information 33 illustrated in FIG.

  Based on the score information 33, the extraction unit 23 determines the region where the cumulative value exceeds a preset threshold value as the deposit region M. Then, the extraction unit 23 outputs the coordinate information in the attachment determination image Ld of the determined attachment region M to the attachment removal device 11 shown in FIG. 2, and instructs the removal operation.

  Thereby, the deposit | attachment removal apparatus 11 will remove the deposit | attachment F adhering to the lens 10a. Further, when instructing such a removal operation, the extraction unit 23 resets the score information 33 to reset the cumulative value of the scores in each region. Note that the extraction unit 23 may reset the score information 33 every predetermined period.

  Thus, the extraction unit 23 can improve the detection accuracy of the deposit F by extracting the deposit region M based on the plurality of time-series adhesion determination images Ld. That is, since the extraction unit 23 does not determine the region extracted as the adhered region M by only one adhesion determination image Ld as the adhered region M, it is possible to suppress erroneous detection of the adhered item F.

  Next, a processing procedure executed by the deposit detection apparatus 1 according to the embodiment will be described with reference to FIG. FIG. 5 is a flowchart illustrating a processing procedure executed by the attached matter detection apparatus 1. Note that this processing procedure is repeatedly executed by the control unit 2 shown in FIG.

  As shown in FIG. 5, the calculation unit 21 calculates the hue value, lightness value, and saturation value of each pixel (step S101). Subsequently, the generation unit 22 generates a binarized image based on the hue value, the lightness value, and the saturation value (step S102).

  Subsequently, the extraction unit 23 extracts an overlapping region by taking a logical product of the binarized images (step S103), and adds a score for each region (step S104). Next, the extraction unit 23 determines whether or not the cumulative value of the score exceeds a threshold value (step S105).

  In this determination, when the cumulative value exceeds the threshold value (Yes at Step S105), the extraction unit 23 determines the region where the cumulative value exceeds the threshold value as the deposit region M (Step S106), and ends the process.

  On the other hand, when the cumulative value is equal to or smaller than the threshold value in the determination in step S105 (No in step S105), the control unit 2 repeats the processing from step 101 onward.

  As described above, the attached matter detection apparatus 1 according to the embodiment includes the calculation unit 21 and the extraction unit 23. The calculation unit 21 calculates the hue value of each pixel included in the captured image L captured by the imaging device (camera 10). Based on the hue distribution that is the hue value distribution calculated by the calculation unit 21, the extraction unit 23 extracts an attachment region M that indicates an imaging region of the attachment F attached to the lens 10a of the imaging device (camera 10). . Therefore, according to the attached matter detection apparatus 1 according to the embodiment, the opaque attached matter F can be detected.

  By the way, in the above-described embodiment, the case where the extraction unit 23 extracts the attachment region M based on the binarized hue image Lh generated by the generation unit 22 has been described. However, the present invention is not limited to this. .

  Therefore, a case where the extraction unit 23 extracts the attachment region M based on the histogram of the hue distribution will be described as a modified example with reference to FIG. FIG. 6 is a diagram illustrating the processing contents of the attached matter detection apparatus 1 according to the modification.

  As shown in FIG. 6, first, the attached matter detection apparatus 1 creates a histogram indicating the hue distribution for each predetermined section (step S21). Subsequently, the adhering matter detection apparatus 1 calculates the similarity between the histogram and the sample histogram (step S22).

  Here, the sample histogram is created in advance by machine learning or the like using, as sample data, a plurality of histograms indicating the hue distributions of the plurality of attachment regions M. That is, the sample histogram is a histogram showing a hue distribution peculiar to the deposit region M.

  Then, the attached matter detection apparatus 1 extracts a region having a high similarity as the attached matter region M (step S23). That is, the attached matter detection apparatus 1 according to the modification detects an opaque attached matter F by extracting a region having a high similarity between the histogram created from the captured image L and the sample histogram as the attached matter region M. be able to.

  Further, for example, the attached matter detection apparatus 1 may calculate the variance value of the histogram, calculate the similarity with the variance value of the sample histogram, and extract the attached matter region M based on the similarity. .

  Further, in the above-described embodiment, the case where the attached matter detection apparatus 1 is applied to the vehicle-mounted camera 10 has been described. However, for example, other monitoring / security cameras set inside or outside the building, alleys, etc. It may be applied to different types of cameras.

  Further effects and modifications can be easily derived by those skilled in the art. Thus, the broader aspects of the present invention are not limited to the specific details and representative embodiments shown and described above. Accordingly, various modifications can be made without departing from the spirit or scope of the general inventive concept as defined by the appended claims and their equivalents.

DESCRIPTION OF SYMBOLS 1 Adhering matter detection apparatus 10 Camera (an example of an imaging device)
10a Lens 21 Calculation unit 22 Generation unit 23 Extraction unit F Attachment L Captured image Lh Binary hue image Lv Binary lightness image Ls Binary saturation image Ld Adhesion determination image M Attachment region

Claims (8)

A calculation unit that calculates a hue value of each pixel included in the captured image captured by the imaging device;
An extraction unit that extracts an attachment region indicating an imaging region of the attachment attached to the lens of the imaging device based on a hue distribution that is a distribution of the hue value calculated by the calculation unit. Attachment detection device. The extraction unit includes:
The attached matter detection apparatus according to claim 1, wherein a region in which the hue value calculated by the calculation unit is a specific color unique to the attached matter is extracted as the attached matter region. The extraction unit includes:
3. The captured image is divided into a plurality of regions, and when the ratio of the specific color in the divided regions satisfies a predetermined condition, the region is extracted as the attached region. The adhering matter detection apparatus as described. The extraction unit includes:
The adhering matter detection apparatus according to claim 3, wherein a peripheral portion of the captured image is divided larger than a portion other than the peripheral portion. The extraction unit includes:
A score is added to the region that satisfies the predetermined condition, and when the cumulative value of the score added to the region for a plurality of time-series captured images is equal to or greater than a threshold value, the region is defined as the attached region. The deposit detection apparatus according to claim 3, wherein the deposit is detected. The calculation unit includes:
Calculating the brightness value of each pixel;
The extraction unit includes:
The attached matter detection apparatus according to claim 2, wherein a region in which the brightness value is less than a threshold value and the hue value is the specific color is extracted as the attached matter region. The calculation unit includes:
Calculating the saturation value of each pixel,
The extraction unit includes:
The deposit according to any one of claims 2 to 6, wherein an area where the saturation value is in a predetermined range and the hue value is the specific color is extracted as the deposit area. Detection device. A calculation step of calculating a hue value of each pixel included in the captured image captured by the imaging device;
An extraction step of extracting an attachment region indicating an imaging region of the attachment attached to the lens of the captured image based on a hue distribution that is a distribution of the hue value calculated by the calculation step. Attachment detection method.

Images