Image recognition-based door-to-door interview method and computer-readable storage medium

Abstract

Disclosed in the present invention are an image recognition-based door-to-door interview method and a computer-readable storage medium. The method comprises: photographing a night scene picture of the outside of a residential building, and transmitting same to a cloud server; the cloud server identifying households with lit windows in the night scene picture of the outside of the residential building and generating corresponding residence information; and the cloud server transmitting the residence information to a communication terminal of a person who conducts on-site interview, such that the person can directly go to interview the households with people at home. The method can effectively obtain residence information of households with people at home in a residential building, and the residence information can be transmitted in real time to a communication terminal of an interviewing person, such that the interviewing person can accurately go to interview the households, thereby reducing and avoiding the situation of being incapable of determining whether residents are at home in spite of multiple times of door knocking.

Description

A home access method based on image recognition and computer-readable storage medium Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a home access method based on image recognition and a computer-readable storage medium. Background Art

[0002] When property management personnel or community management personnel need to visit households or conduct population censuses, economic censuses, etc., they usually need to make an appointment by phone in advance or knock on each door to visit. However, if the household member is not at home, this will cause delays. In order to improve the accuracy and efficiency of household visits, the present invention is proposed.

[0003] Summary of the Invention

[0004] The main technical problem solved by the present invention is that in the existing technology, home visits require prior telephone contact and knocking on the door often encounters the problem that the residents are not at home, which makes the home visit inefficient, time-consuming, and inaccurate. On this basis, a home visit method based on image recognition and a computer-readable storage medium are provided.

[0005] In order to solve the above technical problems, a technical solution adopted by the present invention is to provide a home access method based on image recognition, including the steps of: taking photos of the night view of the exterior of the residential building and transmitting them to a cloud server; the cloud server identifies the residents with bright windows in the night view photos of the exterior of the residential building and generates corresponding residential information; the residential information is transmitted to the communication terminal of the person conducting the on-site visit, so that he can directly access the corresponding residents with bright windows.

[0006] Furthermore, the method for taking photos of the exterior night view of a residential building includes: responding to a trigger signal, entering a building night view shooting mode, first entering a shooting preview stage, displaying the image in the camera on the display screen, automatically identifying the buildings therein, and when there are multiple buildings, framing them with selection boxes to prompt the user to select the building on the display screen; selecting one of the buildings that needs to be visited, entering an automatic focus stage, and the camera achieving fast automatic focus based on the area occupied by the outline of the selected building on the display screen; then entering an automatic shooting stage, automatically adjusting the exposure parameters of the lens for shooting the night view of the building, and obtaining at least two images shot while maintaining the same shooting picture, namely, a first night view building image P1 and a second night view building image P2.

[0007] Furthermore, the first night scene building image P1 corresponds to an image with low exposure parameters, and the overall picture is darker, which is used to accurately identify the bright windows of the building; the second night scene building image P2 corresponds to an image with high exposure parameters, and the overall picture is brighter, which is used to obtain the outer contour of the building and extract lines from it to determine the floor division.

[0008] Furthermore, in the automatic shooting stage, the building is divided into multiple shooting areas according to the lighting conditions of the building to be photographed, and the exposure parameters corresponding to each shooting area are different; wherein, after the first night scene building image P1 is divided and photographed, a first sub-image set {p 11 ,p 12 ,...,p 1n}, where the sub-image p 11 ,p 12 ,...,p 1n The first corrected night scene building image is obtained by merging and stitching the sub-images in the first sub-image set. After the second night scene building image P2 is segmented and photographed, a second sub-image set {p 21 ,p 22 ,...,p 2m}, where the sub-image p 21 ,p 22 ,...,p 2m The second corrected night scene building image is obtained by merging and stitching the sub-images in the second sub-image set.

[0009] Furthermore, the method for the cloud server to identify the residents with bright windows in the night view photo of the exterior of the residential building includes: based on the first night view building image P1 or the first corrected night view building image Perform image recognition to obtain the position information of the bright windows of the building in the image; based on the second night scene building image P2 or the second corrected night scene building image Perform image recognition to obtain the floor information of the building in the image.

[0010] Furthermore, the method for obtaining the position information of the bright windows of the building in the image includes: converting the first night view building image or the first corrected night view building image into a grayscale image, and then performing image binarization processing on the grayscale image, setting the grayscale value of the bright window as the image background of the target identification area to 0, and setting the grayscale value of the bright window in the target identification area to 1, to obtain a two-dimensional matrix with only values ​​0 and 1; traversing the two-dimensional matrix, and whenever a matrix point is found with a grayscale value of 1, identifying and judging the grayscale values ​​of the six adjacent matrix points around the matrix point to determine whether there are connected 1s, and then identifying the set of matrix points occupied by a bright window as an island, and then counting the number and positions of the islands to obtain a bright window identification map.

[0011] Furthermore, the method for obtaining the floor information of the building in the image includes: reading the second night view building image or the second corrected night view building image and converting it into a grayscale image, and using an edge detection operator to convert it into a binary edge image; then performing a Hough transform on the binary edge image; first using a peak detection function to search for peaks in the Hough transform of the binary edge image, and finding Hough transform units greater than a threshold, where the local maximum is a point on the line; identifying a group of candidate peaks, determining line segments related to the candidate peaks and their starting and ending points, corresponding to multiple horizontal and vertical line segments, to obtain a floor identification map.

[0012] Furthermore, the bright window identification map and the floor identification map are superimposed and synthesized to obtain a mixed identification map; each line segment in the mixed identification map is numbered, and the floor and unit location information of the bright window is determined based on the number of the line segment and the number interval range of the line segment where each bright window is located.

[0013] Furthermore, the cloud server also provides a building facade map of the residential building. Based on the building facade map, the number and position of the bright windows in the hybrid identification map are compared to accurately determine whether they belong to the same household or different households.

[0014] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, which stores computer-executable instructions, and the computer-executable instructions are configured to execute and implement the steps of the above-mentioned home access method based on image recognition.

[0015] The beneficial effects of the present invention are as follows: the present invention discloses a method for home access based on image recognition and a computer-readable storage medium. The method comprises: taking a photo of the exterior night view of a residential building and transmitting it to a cloud server; the cloud server identifying the residents with lights on in the photo of the exterior night view of the residential building and generating corresponding residential information; the cloud server then transmits the residential information to the communication terminal of the person conducting the on-site visit, allowing the person to directly access the residents at home. This method can effectively obtain the residential information of residents in the residential building who are at home. This residential information can be transmitted to the visitor's communication terminal in real time, allowing the visitor to conduct precise home access, reducing and avoiding the situation where multiple knocks on the door cannot confirm whether the resident is home. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] FIG1 is a flow chart of an embodiment of a home access method based on image recognition according to the present invention;

[0017] FIG2 is a schematic diagram of an application scenario of an embodiment of a home access method based on image recognition according to the present invention;

[0018] 3 is a schematic diagram of identification and annotation of night view buildings according to another embodiment of the home access method based on image recognition of the present invention;

[0019] 4 is a schematic diagram of a first night view building image according to another embodiment of the home access method based on image recognition of the present invention;

[0020] 5 is a schematic diagram of a second night scene building image according to another embodiment of the home access method based on image recognition of the present invention;

[0021] 6 is a schematic diagram of segmenting a night scene building image into different areas for photographing according to another embodiment of the home access method based on image recognition of the present invention;

[0022] 7 is a schematic diagram of the synthesis of a night scene building image after segmentation and shooting according to another embodiment of the home access method based on image recognition of the present invention;

[0023] 8 is a schematic diagram of binarization processing of a first night scene building image according to another embodiment of the home access method based on image recognition of the present invention;

[0024] 9 is a schematic diagram of bright window recognition after processing and recognizing the first night scene building image according to another embodiment of the home access method based on image recognition of the present invention;

[0025] 10 is a schematic diagram of floor identification after processing and identifying the second night scene building image according to another embodiment of the home access method based on image recognition of the present invention;

[0026] 11 is a schematic diagram showing the synthesis of a bright window identification map and a floor identification map according to another embodiment of the home access method based on image recognition of the present invention;

[0027] FIG12 is a schematic diagram of a building facade that needs to be compared and identified in another embodiment of the home access method based on image recognition according to the present invention. DETAILED DESCRIPTION

[0028] To facilitate understanding of the present invention, the present invention will be described in more detail below with reference to the accompanying drawings and specific embodiments. Preferred embodiments of the present invention are shown in the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described in this specification. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the disclosure of the present invention.

[0029] It should be noted that, unless otherwise defined, all technical and scientific terms used in this specification have the same meanings as those commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are intended solely for the purpose of describing specific embodiments and are not intended to limit the invention. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items.

[0030] FIG1 shows a flow chart of an embodiment of a home access method based on image recognition, including the following steps:

[0031] Step S1: taking photos of the exterior night scene of the residential building and transmitting them to the cloud server;

[0032] Step S2: the cloud server identifies the residents with bright windows in the night view photo of the residential building exterior and generates corresponding residential information;

[0033] Step S3: The residential information is transmitted to the communication terminal of the person conducting the on-site visit, so that the person can directly visit the resident with the corresponding bright window.

[0034] In conjunction with the application scenario diagram shown in FIG2 , the exterior night view photo of residential building 1 is primarily intended to show whether residents have turned on lights. If lights are on, it indicates that someone is home, and the brightness of the windows can be used to distinguish whether lights are on. The photographing terminal 2 that photographs the exterior night view of residential building 1 is typically a user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, vehicle-mounted device, wearable device, etc., with both photographic and communication functions.

[0035] Then, one or more photos of the exterior night scene of the residential building are transmitted to the cloud server 3, which performs intelligent recognition on the photos and identifies the residential information of the residents with lights on, including the floor, unit and house number. Typically, the cloud server belongs to the cloud server of the property management unit or community management unit where the residential building is located.

[0036] Cloud server 3 then transmits the residential information to the communication terminal 4 of the person conducting the on-site visit, enabling the visitor to accurately obtain the information of accessible residences and directly visit the residents at home. The entire process, from taking the photo, transmitting it to the cloud server for identification, to the visitor receiving the information, can be completed within seconds to minutes, offering the advantages of speed and high accuracy.

[0037] As can be seen, the image recognition-based home visit method embodiment shown in Figures 1 and 2 can effectively obtain the residential information of occupied households in a residential building. This residential information can be transmitted in real time to the visitor's communication terminal, allowing them to conduct precise home visits, reducing and avoiding the situation where multiple knocks on the door still cannot confirm whether the resident is home. This application is suitable for community and property management personnel during home visits, especially in applications with specific public management needs, and can greatly improve efficiency and reduce the workload of management personnel.

[0038] Specifically, in step S1, the method for taking photos of the exterior night view of a residential building includes responding to a trigger signal and entering a building night view shooting mode. The viewing system can receive instructions from the processor for display to remind that the current shooting mode is the building night view shooting mode.

[0039] Then, the shooting preview stage is entered, the image in the camera is displayed on the screen, and the buildings in it are automatically identified. When there are multiple buildings, they are framed with selection boxes, prompting you to select the buildings on the display screen.

[0040] As shown in Figure 3, multiple buildings are identified in the picture. Each building has a corresponding feature area, which is surrounded by a rectangular frame and marked as Building. This display frame and label description will also dynamically track as the lens shifts and focal length changes.

[0041] Select one of the buildings that needs to be visited, and enter the autofocus stage. The camera will achieve fast autofocus based on the area occupied by the selected building outline on the display screen, so that the outline of the building can be fully included in the display screen and obtain the largest area and the clearest display effect.

[0042] Specifically, when taking photos, focus on shooting the residential building from the center, that is, shooting from the middle of the building. The height of the shot is also the location of the middle floor of the building. Therefore, it is usually necessary to choose a suitable position opposite the building to shoot. You can also use a drone to shoot, or shoot all around the residential building and then combine multiple photos. However, if the conditions are not met, you can also shoot the residential building from an upward perspective.

[0043] Then, it enters the automatic shooting stage, automatically adjusting the lens's exposure parameters for capturing night scenes of buildings, including automatic adjustment of sensitivity (ISO), shutter speed (S), exposure compensation (EV), focus mode (AF), and white balance (AWB). Automatic adjustment of sensitivity (ISO) is used to set the camera's sensitivity. The higher the ISO sensitivity, the shorter the exposure time, but the more noise in the image. Shutter speed (S) is used to set the exposure time. The device needs to remain still before the exposure time ends, otherwise the image will be blurred. Exposure compensation (EV) is used to change the exposure value recommended by the camera to make the photo brighter or darker. Focus mode (AF) supports single autofocus (AF-S), continuous autofocus (AF-C), and manual focus (MF). Continuous autofocus automatically focuses again when the viewfinder image changes significantly. White balance (AWB) ensures that the image color is not affected by the color of the light source. You can select the corresponding white balance or custom white balance according to the light source type.

[0044] Specifically, in the automatic shooting stage, on the basis of maintaining the same shooting picture, at least two images are captured, namely the first night scene building image P1 and the second night scene building image P2, wherein the first night scene building image P1 corresponds to an image with low exposure parameters, wherein the exposure compensation (EV) is low, and the overall acquired picture is darker, the purpose of which is to obtain a better contrast between bright windows and dark windows, thereby facilitating more accurate identification of lighted windows and residents, as shown in FIG4; the second night scene building image P2 corresponds to an image with high exposure parameters, wherein the exposure compensation (EV) is high, and the overall acquired picture is brighter, the purpose of which is to obtain the outer contour of the building, extract lines from it and determine the floor division, as shown in FIG5.

[0045] Furthermore, during the automatic shooting phase, the building is segmented according to its own lighting conditions, and is divided into multiple shooting areas for separate shooting. The exposure parameters corresponding to each shooting area are different, but the final shooting effect of each shooting area is consistent. Then, these multiple shooting areas are combined into one picture. Therefore, after the first night scene building image P1 is segmented and shot, the first sub-image set {p 11 ,p 12 ,...,p 1n}, the first sub-image {p11 ,p 12 ,...,p 1n} of these sub-images p 11 ,p 12 ,...,p 1n It is taken under different exposure parameters, where n represents the number of these sub-images in the first sub-image set, but the first corrected night scene building image is finally merged and stitched by these sub-images. Similarly, the second night scene building image P2 can also be segmented and photographed to obtain a second sub-image set {p 21 ,p 22 ,...,p 2m}, where m represents the number of these sub-images in the second sub-image set, and the second corrected night scene building image is finally obtained by merging and splicing these sub-images in the second sub-image set.

[0046] Specifically, as shown in Figure 6, since the lower floors of the night scene building are illuminated by external lights, the brightness of the exterior walls is significantly higher than that of the upper floors. Therefore, the building is divided into two areas, upper and lower, and photographed separately, and photographed according to different exposure parameters. Finally, the photos are spliced ​​and synthesized. Figure 7 shows the effect of the actual shooting and splicing.

[0047] Furthermore, the method for the cloud server in step S2 to identify residents with lights on in the exterior night view photo of the residential building includes:

[0048] Based on the first night view building image P1 or the first corrected night view building image Perform image recognition to obtain the location information of the residents with bright windows in the image; based on the second night scene building image P2 or the second corrected night scene building image Perform image recognition to obtain the floor information of the building in the image.

[0049] The method for obtaining the location information of residents with bright windows in an image includes: converting the first night view building image or the first corrected night view building image, converting the original color image into a grayscale image. For example, the original color image is converted into a grayscale image (0-255) based on the RGB value, where white is 0 and black is 255. The grayscale image is then binarized. For example, the average grayscale of the entire image is calculated based on the grayscale value of the current image to serve as the threshold for grayscale segmentation (image binarization). Since the selected image background is basically black or dark gray, the processing is relatively simple. After the image binarization, the grayscale value of the image background of the bright window as the target identification area is set to 0, and the grayscale value of the bright window in the target identification area is set to 1, resulting in a two-dimensional matrix with only values ​​0 and 1, as shown in Figure 8.

[0050] Further preferably, the two-dimensional matrix is ​​traversed, and each time a grayscale value of a matrix point is found to be 1, the grayscale values ​​of the six adjacent matrix points around the matrix point are identified and judged to determine whether there are connected 1s, and then the set of matrix points occupied by a bright window is identified as an island (that is, the pixel points occupied by a bright window), and then the number and position of the islands are counted to obtain a bright window recognition map, as shown in Figure 9.

[0051] Furthermore, the method for obtaining the floor information of the building in the image includes: reading the second night view building image P2 or the second corrected night view building image and converting it into a grayscale image, and using an edge detection operator (such as Canny) to convert it into a binary edge image; then performing a Hough transform on the image; first using a peak detection function to find the peak in the Hough transform of the image, and find the Hough transform unit greater than the threshold. The local maximum value should most likely be a line; for the set of candidate peaks identified above, it is necessary to determine the line segments associated with them and their starting and ending points, such as the horizontal and vertical line segments shown in Figure 10, corresponding to the floor identification map.

[0052] Next, the bright window identification map shown in Figure 9 and the floor identification map shown in Figure 10 are superimposed to form a composite identification map. Each line segment in the composite identification map is numbered. The numbering order can be based on image clarity and perspective, from top to bottom and from near to far. This allows the floor and unit location information of each bright window to be determined based on the line segment number and the range of the line segment number within which the bright window resides.

[0053] Furthermore, to verify and validate the information about residents with bright windows obtained from the hybrid identification map, the cloud server also provides architectural drawings of the relevant residential buildings, including building elevations, as shown in Figure 12. These are actual engineering drawings of the buildings, detailing the location and number of windows for each household. This allows the number and location of bright windows in the hybrid identification map to accurately determine whether they belong to the same household or different households, thus improving the accuracy of household identification.

[0054] Thus, the present invention discloses a method for home access based on image recognition. The method comprises: taking a photo of the exterior night view of a residential building and transmitting it to a cloud server; the cloud server identifying the residents with lights on in the photo of the exterior night view of the residential building and generating corresponding residential information; the cloud server then transmits the residential information to the communication terminal of the person conducting the on-site visit, allowing the person to directly access the residents at home. This method can effectively obtain the residential information of residents in the residential building who are at home. This residential information can be transmitted to the visitor's communication terminal in real time, allowing the visitor to conduct precise home access, reducing and avoiding the situation where multiple knocks on the door cannot confirm whether a resident is home.

[0055] In addition, the present invention also provides a computer-readable storage medium, which stores computer-executable instructions. The computer-executable instructions are configured to execute and implement the steps of the home access method based on image recognition in the above embodiments.

[0056] In the embodiment of the computer-readable storage medium of the present invention, all technical features of the above-mentioned embodiments of the home access method based on image recognition are included. The description and explanation content are basically the same as those of the above-mentioned embodiments of the home access method based on image recognition, and will not be repeated here.

[0057] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, or of course by hardware, but in many cases the former is a more preferred embodiment. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a computer-readable storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above to execute the methods described in various embodiments of the present invention.

[0058] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's description and drawings, or directly or indirectly applied to other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A home access method based on image recognition, characterized in that: Includes steps: Take photos of the exterior night scene of the residential building and transmit them to the cloud server; The cloud server identifies the residents with bright windows in the exterior night view photo of the residential building and generates corresponding residential information; The residential information is transmitted to the communication terminal of the person who conducts the on-site visit, so that he can directly visit the resident of the corresponding bright window.

2. The method for home visit based on image recognition according to claim 1, characterized in that: The method for taking photos of the exterior night scene of a residential building comprises: In response to the trigger signal, the building night scene shooting mode is entered, and the shooting preview stage is first entered. The image in the camera is displayed on the display screen, and the buildings therein are automatically identified. When there are multiple buildings, they are respectively framed with a selection box, prompting the user to select the building on the display screen; Select one of the buildings that needs to be visited, and enter the auto-focus stage. The camera realizes fast auto-focus according to the area occupied by the outline of the selected building on the display screen; Then enter the automatic shooting stage, automatically adjust the exposure parameters of the lens to shoot the night scene of the building, and obtain at least two images shot on the basis of maintaining the same shooting picture, namely the first night scene building image P1 and the second night scene building image P2.

3. The method for home visit based on image recognition according to claim 2, characterized in that: The first night scene building image P1 corresponds to an image with low exposure parameters, and the overall picture is darker, which is used to accurately identify the bright windows of the building; the second night scene building image P2 corresponds to an image with high exposure parameters, and the overall picture is brighter, which is used to obtain the outer contour of the building and extract lines from it to determine the floor division.

4. The method for home visit based on image recognition according to claim 3, characterized in that: In the automatic shooting stage, the building is divided into a plurality of shooting areas for shooting respectively according to the lighting conditions of the building, and each shooting area has a different exposure parameter. The first night scene building image P1 is segmented and photographed to obtain a first sub-image set {p 11 ,p 12 ,...,p 1n }, where the sub-image p 11 ,p 12 ,...,p 1n is taken under different exposure parameters, n represents the number of sub-images in the first sub-image set, and then the sub-images in the first sub-image set are merged and spliced ​​to obtain a first corrected night scene building image After the second night scene building image P2 is segmented and photographed, a second sub-image set {p 21 ,p 22 ,...,p 2m }, where the sub-image p 21 ,p 22 ,...,p 2m is taken under different exposure parameters, m represents the number of sub-images in the second sub-image set, and then the sub-images in the second sub-image set are merged and spliced ​​to obtain a second corrected night scene building image 5. The method for home visit based on image recognition according to claim 4, characterized in that: The method for the cloud server to identify residents with bright windows in the exterior night view photo of the residential building includes: Based on the first night view building image P1 or the first corrected night view building image Perform image recognition to obtain the position information of the bright windows of the building in the image; based on the second night scene building image P2 or the second corrected night scene building image Perform image recognition to obtain the floor information of the building in the image.

6. The method for home visit based on image recognition according to claim 5, characterized in that: The method for obtaining the position information of the bright window of the building in the image comprises: converting the first night view building image or the first corrected night view building image into a grayscale image, then performing image binarization processing on the grayscale image, setting the grayscale value of the bright window as the image background of the target recognition area to 0, and setting the grayscale value of the bright window of the target recognition area to 1, so as to obtain a two-dimensional matrix with only 0 and 1 values; The two-dimensional matrix is ​​traversed, and each time a matrix point with a grayscale value of 1 is found, the grayscale values ​​of the six adjacent matrix points around the matrix point are identified and judged to determine whether there are connected 1s, and then the set of matrix points occupied by a bright window is identified as an island, and then the number and position of the island are counted to obtain a bright window recognition map.

7. The method for home visit based on image recognition according to claim 6, characterized in that: The method for obtaining the floor information of the building in the image comprises: reading the second night view building image or the second corrected night view building image and converting it into a grayscale image, and converting it into a binary edge image using an edge detection operator; then performing Hough transform on the binary edge image; first using a peak detection function to find a peak in the Hough transform of the binary edge image, and finding a Hough transform unit greater than a threshold, where the local maximum is a point on the line; identifying a group of candidate peaks, determining line segments related to the candidate peaks and their starting points and ending points, corresponding to a plurality of horizontal and vertical line segments, and obtaining a floor identification map.

8. The method for home visit based on image recognition according to claim 7, characterized in that: The bright window recognition map and the floor recognition map are superimposed and synthesized to obtain a mixed recognition map; Each line segment in the mixed identification map is numbered, and the floor and unit location information of the bright window is determined according to the line segment number and the number interval range of the line segment where each bright window is located.

9. The method for home visit based on image recognition according to claim 8, characterized in that: The cloud server also provides a building elevation diagram of the residential building. Based on the building elevation diagram, the number and position of bright windows in the mixed recognition image are compared to accurately determine whether they belong to the same household or to different households.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are configured to execute the steps of the home access method based on image recognition as described in any one of claims 1-9.

Images