Image processing device and method and non-transitory computer-readable storage medium
By using multiple different correction quantities in the image processing device for grayscale correction and selecting the larger one, the problem of low estimation accuracy of specific objects under low contrast is solved, and the estimation accuracy is significantly improved.
Patent Information
- Application Number
- CN202010987591.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2019-09-19
- Filing Date
- 2020-09-18
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2040-09-18
AI Technical Summary
The prior art is difficult to accurately estimate the number of specific objects in an image under low contrast conditions, such as in the case of haze or shooting through glass, and traditional grayscale correction methods are not suitable for improving estimation accuracy.
A number of different correction quantities are used to correct the grayscale level of the image, and the larger one is selected by comparing the estimation results after different corrections to improve the estimation accuracy of the number of specific objects. The specific method includes performing grayscale correction using the first correction unit and the second correction unit respectively, and selecting the larger one by the selection unit as the final estimation result.
By emphasizing the contrast of low brightness or high brightness areas, the estimation accuracy of the number of specific objects in the image is significantly improved, especially under low contrast conditions.
Smart Images

Figure CN112529786B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to an image processing device, an image processing method and a non-transitory computer-readable storage medium. Background Art
[0002] There is a known technology for analyzing images acquired at fixed time intervals from a surveillance camera and recording the number of specific objects such as human bodies and vehicles within an image sensing range, with the purpose of grasping a crowded situation or utilizing marketing. Examples of a method of estimating the number of specific objects include a technology for measuring the number of people detected by a human body detection unit, and a technology for estimating the number of specific objects based on a degree of crowding calculated based on a detection result of a change amount and motion between subsequent frame images.
[0003] In the above-mentioned technology, the estimation is performed based on the detection results of feature points and motion calculated from the image, so that the estimation accuracy varies depending on the contrast of the input image and decreases in low-contrast conditions (for example, particularly when the image is captured in fog or mist and when the image is captured through glass).
[0004] As a technique for emphasizing contrast, Japanese Patent No. 6486055 discusses a technique for performing grayscale correction according to a correction degree set from a histogram of an image. Japanese Patent Laid-Open No. 2002-305683 discusses a technique for performing grayscale correction with reference to a specific object area of an image.
[0005] However, since the conventional techniques discussed in the above patent documents are not intended to estimate the number of specific objects, even if the image quality is improved, there may be a situation where the grayscale correction according to the conventional techniques is not preferred in the process of estimating the number of specific objects. Therefore, the conventional techniques are not suitable for improving the estimation accuracy of the number of specific objects. Summary of the invention
[0006] The present invention has been made in view of the above-mentioned problems, and aims to improve the estimation accuracy of the number of specific objects in an image.
[0007] According to a first aspect of the present invention, there is provided an image processing device, comprising: a first correction unit, configured to perform a first grayscale correction on an image using a first correction amount; a first estimation unit, configured to estimate the number of specific objects in the image corrected by the first correction unit; a second correction unit, configured to perform a second grayscale correction on the image using a second correction amount different from the first correction amount; a second estimation unit, configured to estimate the number of specific objects in the image corrected by the second correction unit; and a selection unit, configured to select the larger of the number of specific objects estimated by the first estimation unit and the number of specific objects estimated by the second estimation unit.
[0008] According to a first aspect of the present invention, there is provided an image processing method, comprising: performing a first grayscale correction on an image using a first correction amount; estimating the number of specific objects in the corrected image; performing a second grayscale correction on the image using a second correction amount different from the first correction amount; estimating the number of specific objects in the corrected image; and selecting the larger of the estimated numbers of specific objects.
[0009] According to a first aspect of the present invention, there is provided a non-temporary computer-readable storage medium storing a program, wherein when the program is read and executed by a computer, the computer performs the steps of an image processing method, the image processing method comprising: performing a first grayscale correction on an image using a first correction amount; estimating the number of specific objects in the corrected image; performing a second grayscale correction on the image using a second correction amount different from the first correction amount; estimating the number of specific objects in the corrected image; and selecting the larger of the estimated numbers of specific objects.
[0010] The present invention can improve the estimation accuracy of the number of specific objects in an image.
[0011] Further features of the present invention will become apparent from the following description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] Figure 1 is a block diagram illustrating a configuration example of an image processing apparatus according to a first exemplary embodiment.
[0013] Figure 2 is a flowchart illustrating an example of image processing in the image processing apparatus.
[0014] Figure 3A is a diagram showing an estimated object image, and Figures 3B to 3D is a graph obtained by estimating the object image.
[0015] Figure 4is a block diagram illustrating a configuration example of an image processing system according to a second exemplary embodiment.
[0016] Figure 5 1 is a diagram showing an example of allocation of recorded frame images in which the estimated number of persons is recorded for temporally consecutive frame images by the number recording unit.
[0017] Figure 6 is a flowchart illustrating an example of image processing in the image processing system.
[0018] Figures 7A to 7C are diagrams each showing an example of an estimation object image and position information acquired by a quantity estimation unit. DETAILED DESCRIPTION
[0019] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments are not intended to limit the scope of the claimed invention. A plurality of features are described in the embodiments, but the invention requiring all such features is not limited, and a plurality of such features may be appropriately combined. In addition, in the accompanying drawings, the same reference numerals are assigned to the same or similar structures, and their repeated description is omitted.
[0020] (First Exemplary Embodiment)
[0021] Figure 1 is a block diagram illustrating a configuration example of an image processing apparatus according to a first exemplary embodiment.
[0022] The image processing device includes: an image acquisition unit 1 for acquiring an image; a grayscale correction unit 2 for correcting the grayscale of the acquired image; a quantity estimation unit 3 for estimating the quantity of specific objects in the corrected image; and a quantity recording unit 4 for recording the quantity estimated by the quantity estimation unit 3.
[0023] The image acquisition unit 1 acquires an image from an area (target area) where the number of specific objects is to be estimated. The grayscale correction unit 2 performs grayscale correction on the image acquired by the image acquisition unit 1 based on a predetermined correction amount. The number estimation unit 3 estimates the number of specific objects within the camera range (for example, the number of people in the case where the specific objects are human bodies) by analyzing the image that has been grayscale corrected by the grayscale correction unit 2. The number estimation unit 3 is capable of acquiring position information about the specific objects on the image together with the number of people. The number recording unit 4 records the estimated number of specific objects estimated by the number estimation unit 3. The number recording unit 4 records the position information about the specific objects acquired by the number estimation unit 3.
[0024] The present exemplary embodiment will be described by using an example of estimating the number of human bodies as an example of a specific object, but the specific object is not limited thereto and may be a vehicle, etc. The method of estimating the number of human bodies (number of people) is not particularly limited and may be a method for measuring the number of times a convex shape corresponding to the head to shoulders of a human body is detected, or an estimation method using detection results of other feature quantities or motion quantities.
[0025] Figure 2 is a flowchart illustrating an example of image processing by the information processing apparatus according to the first exemplary embodiment.
[0026] In step S1, the image acquisition unit 1 first acquires an image of the object area. The image acquisition unit 1 may acquire an image from an imaging element such as a complementary metal oxide semiconductor (CMOS) sensor, or acquire an image recorded in an external device such as a hard disk drive. Subsequently, in step S2, the grayscale correction unit 2 generates a plurality of correction curves with different correction amounts. As a method for generating the correction curves, the correction curves may be sequentially generated with reference to a correction curve table prepared in advance, or may be generated by the following (reference Figures 3A to 3D ) describes a method to generate a correction curve corresponding to an image.
[0027] Figures 3A to 3D is a diagram showing a method for generating a calibration curve. Figure 3A shows the estimated object image, and Figures 3B to 3D Each shows an example of a histogram that can be acquired from an estimation object image. Figure 3A 30 is an example of an image 30 which is an estimation object and whose contrast is reduced under the influence of fog or haze. A shaded portion 31 indicates an area where the illumination is locally low, such as a dark area. Figure 3A The histogram of Figure 3B Medium, with high frequency in medium brightness areas.
[0028] The histogram is generally formed by dividing the range of brightness that can be captured by each pixel of the image into a plurality of parts by dividing it by a predetermined number of divisions, and calculating the frequency by counting the number of pixels having brightness belonging to each part by scanning the image. Although an example in which the range of brightness is divided into eight is described for illustration in the present exemplary embodiment, the number of divisions is not specifically limited.
[0029] The following will refer to Figures 3A to 3D Describes an example of a method for generating a calibration curve.
[0030] The grayscale correction unit 2 first divides the image into a plurality of regions according to brightness. Filter processing or the like to reduce high-frequency components of the image may be performed as preprocessing for dividing into regions according to the global brightness distribution.
[0031] A description will be provided of an example in which an image is segmented with a predetermined brightness Y1 used as a boundary.
[0032] For example, in Figure 3A In the example shown, the grayscale correction unit 2 divides the image by setting the area (shaded portion 31) having a brightness less than a predetermined brightness Y1 as a low brightness area and setting the other areas as non-low brightness areas. Subsequently, the grayscale correction unit 2 creates a histogram of the image. At this time, when counting the frequency of pixels present in the low brightness area, the grayscale correction unit 2 does not add 1 to the frequency, but calculates the frequency by adding a value obtained by multiplying 1 by the first weight w1 to the frequency. On the contrary, for the non-low brightness area, the grayscale correction unit 2 calculates the frequency by adding a value obtained by multiplying 1 by the second weight w2 to the frequency. At this time, the grayscale correction unit 2 sets the first weight w1 and the second weight w2 so that the average value of the weights of all pixels becomes 1.
[0033] Figure 3C An example of a weighted histogram created using the weights assigned as described above is shown. If the histogram is calculated using the weights assigned in this manner, then Figure 3C As shown, the resulting histogram indicates that the frequency of low brightness areas is emphasized.
[0034] Thereafter, the grayscale correction unit 2 generates a correction curve 35 having an input / output characteristic inclined in proportion to the frequency in each portion of the weighted histogram, such as Figure 3D As shown. That is, the grayscale correction unit 2 determines the correction curve so that the inclination of the correction curve becomes larger in the area with a higher brightness frequency. The grayscale correction unit 2 can acquire an image that emphasizes the contrast of the low brightness area by performing grayscale correction on the image acquired by the image acquisition unit 1 based on the correction curve created in this way. As a result, the number estimation unit 3 according to the present exemplary embodiment can improve the estimation accuracy of the number in the low brightness area compared to the case where the number is estimated based on the image without grayscale correction. Optionally, by using brightness Y2 higher than brightness Y1 instead of brightness Y1 to divide the image into high brightness areas and non-high brightness areas to create a histogram that emphasizes the frequency in the high brightness area, it is possible to create a correction curve that emphasizes the contrast in the high brightness area. By using the first brightness Y1 and the second brightness Y2, a correction curve that emphasizes the frequency in the medium brightness area with a brightness greater than or equal to the first brightness Y1 and less than or equal to the second brightness Y2 can be generated. In addition, various correction curves can be created by changing the first brightness Y1, the second brightness Y2, the first weight w1, and the second weight w2.
[0035] Return to reference Figure 2In step S3 following step S2, the grayscale correction unit 2 performs grayscale correction based on the generated correction curves and outputs a plurality of images for estimation.
[0036] In step S4, the number estimation unit 3 estimates the number of people for a plurality of images for estimation that have been grayscale corrected by the grayscale correction unit 2. Subsequently, in step S5, the number estimation unit 3 compares the number of people estimated for a plurality of images for estimation. Furthermore, in step S6, the number estimation unit 3 selects the maximum estimated number of people from a plurality of estimated number of people based on the result of the comparison, and provides the maximum estimated number of people to the number recording unit 4. The number recording unit 4 records the estimated number of people provided by the number estimation unit 3. Here, different areas may adopt different maximum estimated number of people in the correction curve, so that the entire area may be pre-divided into a plurality of areas for recording, and the maximum estimated number of people may be recorded for each area for recording. In other words, the number of specific objects included in the image in the object area may be estimated based on the estimated number of object objects in any divided area.
[0037] As described above, in the present exemplary embodiment, the number of specific objects is estimated using an image in which the contrast in a low brightness area is emphasized, which can improve the estimation accuracy of the number of specific objects in the image. In the present exemplary embodiment, the number of people in the corrected image for estimation is estimated, and the maximum estimated number of people is selected from the estimated number of people, which can further improve the estimation accuracy of the number of specific objects in the image.
[0038] (Second embodiment)
[0039] Figure 4 is a block diagram illustrating a configuration example of an image processing system according to a second exemplary embodiment.
[0040] The same constituent elements or corresponding portions of the image processing system as those of the image processing apparatus according to the first exemplary embodiment described above are denoted by the same reference numerals, and description thereof will be omitted.
[0041] The image processing system includes an image pickup apparatus 10 for picking up an image of a subject area and an information processing apparatus 20 for processing the image picked up by the image pickup apparatus 10 .
[0042] The imaging device 10 includes an imaging unit 11 instead of the image acquisition unit 1, a grayscale correction unit 2 for correcting the grayscale of an image from the imaging unit 11, a communication unit 13 for communicating with the information processing device 20, and an estimation information recording unit 14 for recording estimation information.
[0043] The imaging unit 11 outputs continuous frame images captured by imaging to the grayscale correction unit 2. The communication unit 13 transmits the frame images subjected to grayscale correction by the grayscale correction unit 2 to the information processing device 20. The estimation information recording unit 14 records the estimation quantity acquired from the information processing device 20 via the communication unit 13 and the correction curve for the frame image which can be acquired from the grayscale correction unit 2 and used as the estimation object as a set of estimation information.
[0044] The information processing apparatus 20 includes a communication unit 21 for communicating with the camera 10, a video recording unit 22 for recording a video from the camera 10, a setting input unit 23 for inputting settings, a number estimation unit 3, and a number recording unit 4. The video recording unit 22 includes a recording device such as a storage device and a memory device. In the present exemplary embodiment, the number estimation unit 3 is capable of outputting position information indicating the position of an object on an image and an estimated number of people.
[0045] The communication unit 21 transmits the position information estimated by the number estimation unit 3 to the imaging apparatus 10 .
[0046] The video recording unit 22 records the frame images received from the camera 10 via the communication unit 21. The setting input unit 23 inputs confirmation of the frame images recorded in the video recording unit 22, selection of an estimation target area for estimating the number of specific objects, and settings such as the setting of the recording period of the number estimation unit 3 via a user interface such as a display device and a keyboard.
[0047] The following will refer to Figure 5 To describe the periodic number of people estimation in image processing systems.
[0048] Figure 5 2 is a diagram showing the assignment of recorded frame images to temporally continuous frame images. The recorded frame images are frame images in which the estimated number of persons estimated by the number estimation unit 3 is recorded. Figure 5 Recording frame images 40 , 41 , and 42 and non-recording frame images (eg, 43 , 44 , 45 , and 46 ) other than the recording frame images 40 , 41 , and 42 are shown.
[0049] The number estimation unit 3 estimates the number of people in the image in the same manner as in the first exemplary embodiment described above per recording period T. The recording period T can be set using the setting input unit 23. The recording period T can be changed using the setting input unit 23. When the number of frame images provided from the image pickup device 10 during the recording period T is N, there are N-2 non-recording frame images without an estimated number of people between two recording frame images (e.g., between the recording frame images 40 and 41, and between the recording frame images 41 and 42). The present exemplary embodiment uses the N-2 non-recording frame images to search for an appropriate correction curve for the next recording frame image.
[0050] Figure 6 is a flowchart illustrating an example of image processing in the image processing system according to the present exemplary embodiment.
[0051] In step S11, the imaging unit 11 first captures an image of the object area to acquire a frame image. Subsequently, in step S12, the grayscale correction unit 2 determines whether the current frame image from the imaging unit 11 is a recording frame image. In the case where the current frame image is a recording frame image ("Yes" in step S12), the process proceeds to step S13. In step S13, the grayscale correction unit 2 refers to the estimation information recorded in the past in the estimation information recording unit 14 to generate a correction curve representing the maximum estimated number of people.
[0052] If the current frame image is not a recording frame image (No in step S12), the process proceeds to step S14. In step S14, the grayscale correction unit 2 generates a correction curve for searching, and the process proceeds to step S15. A method for generating a correction curve for searching will be described below.
[0053] In step S15 , the grayscale correction unit 2 performs grayscale correction based on the correction curve generated in step S13 or step S14 .
[0054] In step S16 , the number estimation unit 3 estimates the number of people of the frame image acquired from the imaging apparatus 10 via the communication unit 21 .
[0055] Subsequently, in step S17, the number estimation unit 3 determines whether the current frame image is a recording frame image. If the number estimation unit 3 determines that the current frame image is a recording frame image ("Yes" in step S17), the process proceeds to step S18. In step S18, the number estimation unit 3 records the estimated number of people in the number recording unit 4, and the process returns to step S11.
[0056] If the number estimation unit 3 determines that the current frame image is not a recording frame image (No in step S17), the process proceeds to step S19. In step S19, the number estimation unit 3 records the estimated number of people and data related to the correction curve for the current frame image serving as the estimation object as a set of estimation information, and the process returns to step S11.
[0057] Figures 7A to 7C are diagrams each showing an example of an estimation object image and position information acquired by the number estimation unit 3 .
[0058] Reference below Figures 7A to 7C A method for generating a calibration curve for searching will be described.
[0059] Although in the method of generating a correction curve for searching, a weighted histogram is generated in the same manner as in the first embodiment to generate a correction curve, in the second embodiment, weights are generated using position information recorded as estimation information. A description will be provided of a case where an image is divided into 6×4 rectangular blocks and weights for each block are calculated, for example, as in Fig. 7A For example, the weight W is calculated by the following expressions (1) and (2): (i,j) .
[0060]
[0061]
[0062] In expression (1), N (x,y) is the estimated number of people in a block, N sum is the estimated number of people in the entire frame image, such as Figure 7B In expression (1), σ is an arbitrary coefficient, and n and m are the number of region divisions. In this embodiment, n=6 and m=4. Figure 7C An example of weights calculated using σ = 1 is shown. Figure 7C It can be seen that using expressions (1) and (2) to calculate weights can make the weights larger in blocks with higher human density. Here, since the quantity estimation unit 3 is arranged in a later stage than the grayscale correction unit 2, the position information estimated for the frame image before the current frame image is used. Therefore, the update rate of the position information is likely to be reduced due to the delay of the communication units 13 and 21 and according to the calculation cost (computation time) of the quantity estimation unit 3. In order to avoid this situation, instead of all the segmented areas, every other segmented area in the rectangular block-shaped segmented area can be used as the estimation object area of the quantity estimation unit 3. For the non-estimation object area, the result of linear prediction of the estimated quantity in the estimation object area around it can be used as the estimated quantity.
[0063] As described above, processing is performed according to density, which emphasizes contrast by focusing on areas where the possibility of human presence is high, thereby increasing the efficiency of estimating the number of specific objects in the image. In contrast, the possibility of the emergence of new specific objects or the possibility of the existence of specific objects that were not detected by the estimation in the initial stage of the number estimation unit 3 can be considered. Specifically, in the case where the estimated number does not change or decreases between frames for a predetermined number of frames, a weighted calculation can be performed so that the weight of the area with low density is increased. The grayscale correction unit 2 can perform grayscale correction on a region-by-region basis. In this case, the grayscale correction unit 2 can update the correction curve only for the area where the estimated number increases on a region-by-region basis, and independently perform grayscale correction on the separate areas.
[0064] Since different grayscale corrections are performed for each frame for the image stored in the video recording unit 22, a sharp change in brightness may occur between adjacent frame images. Therefore, the grayscale correction unit 2 can use the weight calculated for the current frame image, the correction curve and the weight calculated for the immediately previous frame image, or a correction curve synthesized using a predetermined coefficient. At this time, in order to respond to the sharp changes in viewing angle and illumination, etc., an image change amount calculation unit (not shown) that can calculate the amount of change and / or motion between frames using background subtraction can be added. In this case, the grayscale correction unit 2 can be set so that the larger the amount of change between frames, the larger the weight calculated for the current frame image using a predetermined coefficient becomes, or the greater the influence of the correction curve becomes.
[0065] In the case where the estimation target region is selected by the setting input unit 23, the calculation cost can be reduced by limiting the region to be subjected to gray scale correction, under the condition that the estimation target region is included in the processing target region in steps S12 to S19.
[0066] As described above, the present exemplary embodiment can improve the estimation accuracy of the number of specific objects in an image as in the first exemplary embodiment. In the present exemplary embodiment, searching for a correction curve suitable for the next recorded frame image using a non-recorded frame can further improve the estimation accuracy of the number of specific objects in an image. According to the present exemplary embodiment, determining the correction amount according to the density of the specific object, and determining the correction amount so as to emphasize the brightness frequency in an area where the density is higher than a threshold value can further improve the estimation accuracy of the number of specific objects in an image.
[0067] Although the desired typical embodiments of the present invention have been described, the present invention is not limited to these typical embodiments, and various modifications and changes may be made without departing from the gist of the present invention. Figure 4The gray level correction unit 2 is arranged in the imaging apparatus 10, but the gray level correction unit 2 may be arranged in the information processing apparatus 20. Although the imaging apparatus 10 and the information processing apparatus 20 are shown as separate apparatuses, the information processing apparatus 20 may be incorporated in the imaging apparatus 10.
[0068] (Other Typical Embodiments)
[0069] Although the exemplary embodiments have been described in detail above, the present invention may be applicable to exemplary embodiments such as systems, devices, methods, programs, or recording media (storage media). More specifically, the present invention may be applicable to a system including a plurality of devices (e.g., a host computer, an interface device, a camera device, and a web application), or may be applicable to a device including a single device.
[0070] The embodiments of the present invention may also be implemented by providing software (program) for performing the functions of the above-described embodiments to a system or device via a network or various storage media, and a computer or a central processing unit (CPU) or a microprocessing unit (MPU) of the system or device reads and executes the program.
[0071] While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments.The scope of the following claims is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions.
Claims
1. An image processing device, comprising: A first correction unit configured to perform a first grayscale correction on the image using a first correction amount; a first estimating unit configured to estimate the number of specific objects in the image corrected by the first correcting unit; a second correction unit configured to perform a second grayscale correction on the image using a second correction amount different from the first correction amount; a second estimating unit configured to estimate the number of specific objects in the image corrected by the second correcting unit; as well as a selecting unit configured to select a larger one of the number of specific objects estimated by the first estimating unit and the number of specific objects estimated by the second estimating unit, wherein the first correction unit determines the first correction amount so that the first correction amount is larger in a region having a higher brightness frequency, and The second correction unit determines the second correction amount so that the second correction amount is larger in a region with a higher brightness frequency.
2. The image processing device according to claim 1, further comprising a calculation unit configured to calculate a histogram of brightness of the image; in, The first correction unit determines the first correction amount using the histogram calculated by the calculation unit, and The second correction unit determines the second correction amount using the histogram calculated by the calculation unit.
3. The image processing device according to claim 2, in, The first correction unit assigns weights so as to emphasize the brightness frequency of the area set based on the first brightness to determine the first correction amount, and The second correction unit determines the second correction amount by assigning weights so as to emphasize the brightness frequency of the area set based on the second brightness.
4. The image processing apparatus according to claim 1 , further comprising a segmentation unit configured to segment the image of the object area into a first area and a second area, in, The first correction unit performs the first grayscale correction on the first area, and the second correction unit performs the second grayscale correction on the second area.
5. The image processing device according to claim 4 further includes a third estimation unit, which is configured to estimate the number of specific objects included in the image based on the number of specific objects estimated by the first estimation unit or the second estimation unit in any one of the first area and the second area into which the image is segmented by the segmentation unit.
6. The image processing device according to claim 1, in, The image includes a plurality of frame images. The first correction unit performs the first grayscale correction on each frame image of the plurality of frame images. The second correction unit performs the second grayscale correction on each frame image of the plurality of frame images. Wherein, the image processing device further includes: a first synthesis unit configured to synthesize a first correction amount for each of the plurality of frame images; and A second synthesis unit is configured to synthesize a second correction amount for each of the plurality of frame images.
7. The image processing device according to claim 6, further comprising: A second detection unit, configured to detect a change amount between the plurality of frame images; as well as A changing unit is configured to change the first correction amount or the second correction amount according to the change amount detected by the second detection unit.
8. The image processing device according to claim 1, further comprising: an input unit configured to input an area in which the first estimation unit estimates the number of specific objects, or an area in which the second estimation unit estimates the number of specific objects; as well as A control unit configured to control a region to be subjected to the first grayscale correction by the first correction unit or to be subjected to the second grayscale correction by the second correction unit.
9. An image processing device, comprising: A first correction unit configured to perform a first grayscale correction on the image using a first correction amount; a first estimating unit configured to estimate the number of specific objects in the image corrected by the first correcting unit; a second correction unit configured to perform a second grayscale correction on the image using a second correction amount different from the first correction amount; a second estimating unit configured to estimate the number of specific objects in the image corrected by the second correcting unit; as well as a selecting unit configured to select a larger one of the number of specific objects estimated by the first estimating unit and the number of specific objects estimated by the second estimating unit, wherein the first correction unit determines the first correction amount so as to emphasize the brightness frequency of an area where the density of the specific object is higher than a threshold value, and The second correction unit determines the second correction amount so as to emphasize the brightness frequency of a region where the density of the specific object is higher than a threshold.
10. The image processing apparatus according to claim 9, further comprising a segmentation unit configured to segment the image of the object area into a first area and a second area, in, The first correction unit performs the first grayscale correction on the first area, and the second correction unit performs the second grayscale correction on the second area.
11. The image processing device according to claim 10 further includes a third estimation unit, which is configured to estimate the number of specific objects included in the image based on the number of specific objects estimated by the first estimation unit or the second estimation unit in any one of the first area and the second area into which the image is segmented by the segmentation unit.
12. The image processing device according to claim 9, further comprising a first detection unit configured to detect the density of the specific object, in, The first correction unit determines the first correction amount according to the density of the specific object detected by the first detection unit. The second correction unit determines the second correction amount according to the density of the specific object detected by the first detection unit.
13. The image processing device according to claim 12, in, The first correction unit determines the first correction amount so as to emphasize the brightness frequency of an area where the density of the specific object detected by the first detection unit is higher than a threshold value, and The second correction unit determines the second correction amount so as to emphasize the brightness frequency of an area where the density of the specific object detected by the first detection unit is higher than a threshold.
14. The image processing device according to claim 9, in, The image includes a plurality of frame images. The first correction unit performs the first grayscale correction on each frame image of the plurality of frame images. The second correction unit performs the second grayscale correction on each frame image of the plurality of frame images. Wherein, the image processing device further includes: a first synthesis unit configured to synthesize a first correction amount for each of the plurality of frame images; and A second synthesis unit is configured to synthesize a second correction amount for each of the plurality of frame images.
15. The image processing device according to claim 14, further comprising: A second detection unit, configured to detect a change amount between the plurality of frame images; as well as A changing unit is configured to change the first correction amount or the second correction amount according to the change amount detected by the second detection unit.
16. The image processing device according to claim 9, further comprising: an input unit configured to input an area in which the first estimation unit estimates the number of specific objects, or an area in which the second estimation unit estimates the number of specific objects; as well as A control unit configured to control a region to be subjected to the first grayscale correction by the first correction unit or to be subjected to the second grayscale correction by the second correction unit.
17. An image processing method, comprising: Performing a first grayscale correction on the image using a first correction amount; estimating the number of specific objects in the rectified image; Performing a second grayscale correction on the image using a second correction amount different from the first correction amount; estimating the number of specific objects in the rectified image; as well as Select the larger of the estimated quantities of a particular object, wherein the first correction amount is determined so that the first correction amount is larger in a region having a higher brightness frequency, and The second correction amount is determined so that the second correction amount is larger in a region with a higher brightness frequency.
18. An image processing method, comprising: Performing a first grayscale correction on the image using a first correction amount; estimating the number of specific objects in the rectified image; Performing a second grayscale correction on the image using a second correction amount different from the first correction amount; estimating the number of specific objects in the rectified image; as well as Select the larger of the estimated quantities of a particular object, wherein the first correction amount is determined so as to emphasize the brightness frequency of an area where the density of the specific object is higher than a threshold, and The second correction amount is determined so as to emphasize the brightness frequency of a region where the density of the specific object is higher than a threshold.
19. A non-transitory computer-readable storage medium storing a program, wherein when the program is read and executed by a computer, the computer performs the steps of an image processing method, the image processing method comprising: Performing a first grayscale correction on the image using a first correction amount; estimating the number of specific objects in the rectified image; Performing a second grayscale correction on the image using a second correction amount different from the first correction amount; estimating the number of specific objects in the rectified image; as well as Select the larger of the estimated quantities of a particular object, wherein the first correction amount is determined so that the first correction amount is larger in a region having a higher brightness frequency, and The second correction amount is determined so that the second correction amount is larger in a region with a higher brightness frequency.
20. A non-transitory computer-readable storage medium storing a program, wherein when the program is read and executed by a computer, the computer performs the steps of an image processing method, the image processing method comprising: Performing a first grayscale correction on the image using a first correction amount; estimating the number of specific objects in the rectified image; Performing a second grayscale correction on the image using a second correction amount different from the first correction amount; estimating the number of specific objects in the rectified image; as well as Select the larger of the estimated quantities of a particular object, wherein the first correction amount is determined so as to emphasize the brightness frequency of an area where the density of the specific object is higher than a threshold, and The second correction amount is determined so as to emphasize the brightness frequency of a region where the density of the specific object is higher than a threshold.
Citation Information
Patent Citations
JP1989086055A
Image pickup device and recording medium for recording image pickup program
JP2002305683A
Real-time face tracking in a digital image acquisition device
CN101512549A
Detection method and device of touch points for image recognition
CN102129332A
