An image generation method and device, a method for removing interference in an image, and a cooker

Abstract

The present application can provide an image generation method and device, an image interference removal method and a cooker, wherein the image generation method comprises one or more steps as follows. An interference-free image is collected, and the interference-free image is an image of a real scene. The interference-free image is subjected to grid division processing to cut the interference-free image into a plurality of first sub-images. A second sub-image is selected from the plurality of first sub-images, and the remaining first sub-images are saved. The second sub-image is subjected to interference processing to generate a third sub-image. The third sub-image is combined with the remaining first sub-images to generate an interference image. The present application can generate a large number of interference images corresponding to the interference-free image under the premise of low cost, the whole process does not need manual intervention, and provides an original data set for training of a generative adversarial network model. It can be seen that the present application has the advantages of high intelligence and low cost.

Description

Technical Field

[0001] This invention relates to the field of intelligent cooking appliances technology. More specifically, this invention can provide an image generation method and apparatus, a method for removing interference from images, and a cooking appliance. Background Technology

[0002] With the development of the information society, networked and smart home appliances are increasingly appearing in people's lives, and the fundamental goal of all this development is to provide users with a smart, comfortable, and convenient home environment. Among these, the smart kitchen is a crucial component of home intelligence and comfort, and smart cookware is paramount. Some have proposed applying artificial intelligence (AI) technology to smart kitchens to accelerate their development. However, due to the complexity of the kitchen environment, such as the effects of fumes, water mist, grease, and water droplets, the information required for AI technology to collect is highly complex. Image information, in particular, is most severely affected, resulting in a small number and poor quality of datasets needed for AI technology. Furthermore, this method of obtaining samples is costly. Especially, as users use their kitchens for extended periods, the kitchen environment will become even more complex. Summary of the Invention

[0003] One of the main objectives of this invention is to provide an image generation method and apparatus, an image interference removal method, and a cooking utensil, in order to solve the problems of insufficient sample quantity and high sample acquisition cost in conventional solutions, thereby achieving technical objectives such as generating non-uniform images with interference and reducing investment costs.

[0004] To address the aforementioned technical problems, the present invention provides an image generation method, which includes, but is not limited to, one or more of the following steps: Acquiring an interference-free image, wherein the interference-free image is an image of a real scene. Dividing the interference-free image into a grid to segment it into multiple first sub-images. Selecting a certain number of second sub-images from the multiple first sub-images and saving the remaining first sub-images. Applying interference processing to these second sub-images to generate a third sub-image. Combining the third sub-image with the remaining first sub-images to generate an image with interference.

[0005] To address the aforementioned technical problems, this invention provides a method for removing interference from images; this method includes, but is not limited to, at least one of the following steps: Obtaining an interference-free image, and generating an image with interference based on the interference-free image using the image generation method described in this embodiment of the invention. Forming an image pair training set by combining multiple pairs of interference-free and interference-interference images, and training an adversarial generative network (GAN) model using this image pair training set. Finally, using the trained GAN model to remove interference from the image to be processed, obtaining an image that almost completely approximates the real scene.

[0006] To address the aforementioned technical problems, the present invention provides an image generation apparatus, which includes, but is not limited to, an image acquisition module, a grid division module, a sub-image selection module, a sub-image interference module, and a sub-image combination module. The image acquisition module acquires an interference-free image, which is an image of a real scene. The grid division module performs grid division processing on the interference-free image to divide it into multiple first sub-images. The sub-image selection module selects a second sub-image from the multiple first sub-images and saves the remaining first sub-images. The sub-image interference module performs interference processing on the second sub-images to generate a third sub-image; the sub-image combination module combines the third sub-image with the remaining first sub-images in an orderly manner to generate an image with interference.

[0007] To address the aforementioned technical problems, the present invention also provides an intelligent cooking appliance, which may include, but is not limited to, the image generation device described in the embodiments of the present invention.

[0008] To address the aforementioned technical problems, the present invention also provides a computer storage medium. This computer storage medium stores an image generation program, and when the image generation program is executed by a processor, it implements the image generation method of any embodiment of the present invention.

[0009] The beneficial effects of this invention are as follows: This invention can generate a large number of disturbed images corresponding to disturb-free images at low cost, without requiring manual intervention, thus providing a raw dataset for training adversarial generative network models. Therefore, this invention has outstanding advantages such as high intelligence and low cost.

[0010] This invention significantly reduces the difficulty of acquiring massive amounts of image samples, generating images with non-uniform interference that more closely resemble real-world scene images (such as images of cooking with oil fumes or watermarked dishwasher images). This results in higher accuracy for the intelligent algorithms applied with this invention. Therefore, this invention can significantly improve the generalization, robustness, and applicability of the model, greatly enhancing the recognition accuracy of image processing algorithms, and ultimately providing users with a better interactive experience in smart kitchens and intelligent cooking.

[0011] Based on the technical solution provided by the embodiments of the present invention, interference information in images can be effectively removed. Taking cooking images of stir-frying as an example, the present invention can remove oil fume interference from cooking images; when washing dishes in a dishwasher, for the captured images, the present invention can remove water stains, foam, and other interference from the images. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0013] Figure 1 A schematic flowchart of a method for generating an interfering image from an interference-free image is shown in an embodiment of the present invention.

[0014] Figure 2 This diagram illustrates a method for dividing an interference-free image into grids using M×N grids, as shown in an embodiment of the present invention.

[0015] Figure 3 This illustration shows a schematic diagram of selecting K second sub-images from M×N first sub-images using a random distribution method in an embodiment of the present invention.

[0016] Figure 4 This illustration shows a schematic diagram of selecting K second sub-images from M×N first sub-images using a connected region approach in an embodiment of the present invention.

[0017] Figure 5 This illustration shows a schematic diagram of selecting K second sub-images from M×N first sub-images using a centrally connected region method in an embodiment of the present invention.

[0018] Figure 6 This illustration shows a schematic diagram of the algorithm flow for selecting a second sub-image from a first sub-image using a grid selection method in an embodiment of the present invention.

[0019] Figure 7 The diagram illustrates the algorithm flow for generating multiple pairs of images with and without interference from a single interference-free image in an embodiment of the present invention.

[0020] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0022] In view of the problems of small number of image sample sets with and without interference and high cost of sample acquisition, the present invention provides an image generation method and apparatus, an image interference removal method, and a cookware, which can effectively solve at least one problem of conventional technology.

[0023] like Figure 1 As shown, embodiments of the present invention can provide an image generation method, which includes, but is not limited to, one or more of the following steps.

[0024] First, this invention requires the acquisition of interference-free images, which are images of real-world scenes. In this embodiment of the invention, interference-free images can be images of kitchen scenes, such as images without oil fumes or watermarks, but are not limited to these.

[0025] The process of acquiring interference-free images in this embodiment of the invention may include: providing a kitchen environment with a clear field of view, free from image interference factors such as oil fumes, water mist, oil stains, or water droplets. A camera mounted on a cooking appliance is then activated in this kitchen environment to capture an image of the current real-world scene, which serves as the interference-free image. The cooking appliance can be, for example, a range hood or dishwasher. Thus, this invention enables the deployment of a camera on a cooking appliance such as a range hood or dishwasher, allowing the camera to capture images (i.e., pictures) of the corresponding scene. The image information can then be processed using artificial intelligence methods.

[0026] It is understood that this embodiment captures kitchen scenes without any interference, resulting in high-quality images of the kitchen-related scenes (i.e., interference-free images acquired in this embodiment).

[0027] like Figure 7 As shown, this invention can save multiple interference-free images and construct an interference-free dataset using these saved images. When preparing to process interference-free images, embodiments of this invention can obtain a single interference-free image from the interference-free dataset using a non-repeating selection method for grid partitioning. It is understood that embodiments of this invention can obtain one or more interference-containing images from a single interference-free image, achieved by setting an iteration round L. Here, L is a natural number, representing that L irregular interference-containing images can be generated from a single interference-free image. Embodiments of this invention do not limit the specific value of L; it can be reasonably set according to actual conditions based on this invention.

[0028] Secondly, after obtaining the interference-free image, the interference-free image is divided into grids to divide it into multiple first sub-images. It can be seen that the embodiments of the present invention can provide a grid-based method for generating interference images.

[0029] Specifically, in this embodiment of the invention, a triplet of natural numbers (M, N, K) is used to represent the number of grid divisions and the number of grid samples. Here, M represents the number of rows in the grid, N represents the number of columns in the grid, and K represents the number of grids to be sampled from M×N grids, i.e., the number K of the second sub-images to be determined.

[0030] Optionally, M represents the number of equally spaced rows divided according to the image height, and N represents the number of equally spaced columns divided according to the image width. The total number of grid cells obtained after division is then M × N, for example... Figure 2 The diagram is shown in the image.

[0031] Next, a preset number of second sub-images are selected from multiple first sub-images, and the remaining first sub-images are saved. This embodiment of the invention uses a grid sampling method to determine the first sub-image corresponding to the sampled grid, thus achieving the selection of the second sub-image.

[0032] Where M×N = the preset number of second sub-images + the remaining number of first sub-images.

[0033] like Figure 3 As shown, optionally, selecting a second sub-image from multiple first sub-images includes: traversing all first sub-images, i.e., M×N sub-images, and calculating the probability of interference for each first sub-image according to a probability distribution; for example, the probability of being affected by cooking fumes, the probability of having water stains, etc. The probability distribution in this embodiment includes either a Gaussian distribution or a random distribution. Based on the probability of interference, K second sub-images are selected from all first sub-images: specifically, K grids are selected from an M×N grid to achieve the selection of K second sub-images from the M×N first sub-images.

[0034] like Figure 4 As shown, and can be combined Figure 6 In this embodiment of the invention, selecting a second sub-image from multiple first sub-images may include: randomly selecting a grid from M×N grids, i.e., randomly selecting a first sub-image as the second sub-image; and, according to connected component analysis, continuing to select first sub-images from the adjacent first sub-images of the current second sub-image as new second sub-images. Optionally, this embodiment of the invention may return to the step of continuing to select new second sub-images based on the current number of second sub-images, or end the second sub-image selection step. Specifically, if the current number of second sub-images is less than K, then the selection and judgment continue from the adjacent first sub-images of the current second sub-image; if the current number of second sub-images is equal to or greater than K, then the sub-image selection step ends, i.e., the loop terminates.

[0035] Optionally, the process of selecting a new second sub-image using connected component analysis can include: obtaining the pixel values ​​of each first sub-image adjacent to the current second sub-image; calculating the pixel difference between each first sub-image and the current second sub-image; and selecting the corresponding first sub-image as the new second sub-image based on the pixel difference. For example, the adjacent first sub-image with the smallest pixel difference can be selected as the second sub-image; or all first sub-images with a pixel difference less than or equal to a preset value can be selected as the second sub-image. It is understood that in this embodiment, "adjacent" between two sub-images means that the grids containing the two sub-images are directly connected or share a common edge.

[0036] like Figure 5 As shown, the grid division processing of the interference-free image in this invention can also include: setting an odd number of rows and columns of grids in the interference-free image area, and performing equidistant grid division processing on the interference-free image. M is odd and N is odd to ensure that there is a central grid in the M×N grids. The position of the central grid can be understood as ((M+1) / 2, (N+1) / 2), that is, the sub-image corresponding to the central grid can be determined. Then, the selection of the second sub-image from multiple first sub-images in this embodiment of the invention includes: taking the first sub-image located at the center of the interference-free image, i.e., the position ((M+1) / 2, (N+1) / 2), as the second sub-image; continuing to select a new second sub-image from the first sub-images adjacent to the current second sub-image according to the central connected region analysis method; and then returning to the step of continuing to select a new second sub-image according to the current number of second sub-images, or ending the second sub-image selection step. If the number of second sub-images is less than K, the selection and judgment will continue in the first sub-images adjacent to the current second sub-image; if the number of second sub-images is equal to or greater than K, the second sub-image selection step will end, that is, the loop process will terminate.

[0037] After determining a suitable number K of second sub-images, the second sub-images are subjected to interference processing to generate third sub-images. Specifically, the third sub-images are obtained by forming or setting interference information on the second sub-images; that is, one third sub-image corresponds to one second sub-image.

[0038] Therefore, this invention can sample the grid using probability distribution or connected component analysis, that is, select some sub-images from the first sub-image as the second sub-image. The result of these sampling methods is non-uniform, meaning the interference generated in the image by this invention is non-uniform. Compared to uniform interference, the solution provided by this invention is more suitable for interference from oil fumes, water mist, or oil stains in a kitchen environment. Therefore, this invention can ultimately produce an image with interference that is almost identical to a real-world image (e.g., an image of a cooking scene while stir-frying food, or a scene of washing dishes in a dishwasher).

[0039] Finally, the third sub-image is combined with the remaining first sub-image to generate one or more disturbed images. Specifically, since there is a one-to-one correspondence between the third sub-image, the second sub-image, and the selected first sub-image, embodiments of the present invention can re-synthesize the third sub-image according to the original sub-image arrangement order to create a new image, i.e., a non-uniform disturbed image. It can be seen that the disturbed image of the present invention originates from and corresponds to the non-disturbed image. Moreover, the disturbance in the disturbed image generated by the present invention is non-uniform, more closely resembling the image in a real scene, and the image size is the same as the original non-disturbed image.

[0040] like Figure 7 As shown, the image generation method further includes: corresponding to the aforementioned iteration round L, this embodiment of the invention determines whether the number of interfering images generated from the current interference-free image has reached iteration round L; if iteration round L has not been reached, the grid partitioning parameters are reselected, and the step of grid partitioning the current interference-free image is returned; the grid partitioning parameters include at least one of the number of rows, the number of columns, and the number of second sub-images to be determined. The process of generating a new interfering image from the current interference-free image ends when iteration round L is reached, i.e., the loop terminates.

[0041] Based on the same technical concept as the image generation method of the present invention, embodiments of the present invention can also provide an image generation device, which may include, but is not limited to, an image acquisition module, a grid division module, a sub-image selection module, a sub-image interference module, and a sub-image combination module.

[0042] The image acquisition module is used to acquire interference-free images, which are images of real-world scenes. This invention's image acquisition module can also be used to construct an interference-free dataset from multiple interference-free images, and can be used to obtain an interference-free image from this dataset using a non-repeating selection method.

[0043] Optionally, the image acquisition module can be used to activate the camera mounted on the cookware in a kitchen environment with a clear view, thereby controlling the camera to acquire images of the current real scene and using the acquired images as interference-free images.

[0044] The grid division module is used to perform grid division processing on the acquired interference-free image to divide the interference-free image into multiple first sub-images.

[0045] The sub-image selection module is used to select a second sub-image from multiple first sub-images and save the remaining first sub-images.

[0046] Optionally, the sub-image selection module is used to traverse all first sub-images of the current interference-free image, and can be used to calculate the probability of interference for each first sub-image according to a probability distribution, and to select one or more second sub-images from all first sub-images based on the obtained probability of interference. The probability distribution in this embodiment of the invention includes one of Gaussian distribution, random distribution, etc.

[0047] Optionally, the sub-image selection module is used to randomly select a first sub-image as the second sub-image, and to use connected component analysis to select new second sub-images from the adjacent first sub-images of the currently selected second sub-image. The sub-image selection module is also used to determine the current number of second sub-images, and to continue selecting new second sub-images if the determination result is less than K, or to end the second sub-image selection step if the determination result is equal to or greater than K. Further, the sub-image selection module is used to obtain the pixel values ​​of each first sub-pixel adjacent to the current second sub-image, and to calculate the pixel difference between the pixel values ​​of these sub-pixels and the pixel value of the current second sub-image, and to select the corresponding first sub-image as the new second sub-image if the pixel difference is less than a preset value.

[0048] Optionally, the sub-image selection module of the present invention is used to select a first sub-image located at the center of the interference-free image as a second sub-image, and based on this, uses the central connected region analysis method to continue selecting new second sub-images from the first sub-images adjacent to the currently selected second sub-image. The sub-image selection module is also used to determine the current number of second sub-images, and continue selecting new second sub-images if the determination result is less than K, or to end the second sub-image selection step if the determination result is equal to or greater than K. Further, the sub-image selection module is used to obtain the pixel values ​​of each first sub-pixel adjacent to the current second sub-image, and to calculate the pixel difference between the pixel values ​​of these sub-pixels and the pixel values ​​of the current second sub-image, and to select the corresponding first sub-image as a new second sub-image if the pixel difference is less than a preset value. It is understood that in this embodiment, the grid division module is used to set an odd number of rows and odd number of columns of grid on the interference-free image area, and to perform equidistant grid division processing on the interference-free image.

[0049] The sub-image interference module is used to interfere with the second sub-image to generate the third sub-image.

[0050] The sub-image combination module is used to combine the third sub-image with the remaining first sub-image to generate an image with interference.

[0051] The image generation apparatus of the present invention may further include an iteration judgment module. The iteration judgment module is used to determine whether the number of interfering images generated from the current interference-free image has reached a preset value; and is used to reselect grid partitioning parameters based on the condition that the iteration round L has not been reached, and to control the grid partitioning module to use the reselected grid partitioning parameters to perform grid partitioning on the current interference-free image, thereby generating new interfering images. Alternatively, it is used to terminate the processing of the current interference-free image based on the condition that the iteration round L has been reached. The grid partitioning parameters may include, but are not limited to, the number of grid rows, the number of grid columns, and the number of second sub-images.

[0052] It is understood that the image generation apparatus provided by this invention can be integrated into a processing chip. This processing chip can integrate basic image operation software libraries and multi-threaded processing libraries to accelerate the overall image processing speed.

[0053] Based on the image generation method of the present invention, the present invention can also provide a method for removing interference in an image, which may include, but is not limited to, at least one of the following steps.

[0054] First, obtain an interference-free image.

[0055] Next, an interfering image is generated based on the acquired interference-free image; the interfering image is obtained through the image generation method in any embodiment of the present invention.

[0056] Secondly, an image pair training set is formed using corresponding disturbed and undisturbed images, and the adversarial generative network model is trained using this training set. The image pair training set consists of disturbed and undisturbed image pairs, with each pair including a disturbed image and its corresponding undisturbed image.

[0057] Finally, the trained generative adversarial network model is used to remove interference from the image to be processed, so as to obtain a real interference-free image.

[0058] When applied to artificial intelligence tasks, this invention can remove interference from the currently acquired image. Even if the acquired image contains a large amount of oil fumes, water mist, and / or oil stains, water droplets, or other interference on the lens, this invention can effectively remove them, improving product quality and user experience.

[0059] In particular, even if different cooking habits and cooking environments lead to different cooking scene images (such as images of cooking fumes), the embodiments of the present invention can still realistically reproduce the real scene based on the adversarial generative network model trained on the interfering images.

[0060] Based on the same inventive concept as the image generation method in the embodiments of the present invention, the present invention can also provide a smart cooker, which may include, but is not limited to, the image generation device in any embodiment of the present invention. Optionally, the smart cooker is a smart range hood or dishwasher, etc.

[0061] Therefore, this invention provides a smart range hood with image generation and interference removal capabilities. In this embodiment, a camera is deployed on the smart range hood to acquire interference-free images. This invention can generate multiple images with non-uniform interference features based on the interference-free images, thereby constructing a "interference + interference-free" dataset, which can then be used for efficient training of an interference removal model (generative adversarial network model). This invention allows the trained interference removal model to be deployed on smart cooking appliances, such as within the processor of a smart range hood or dishwasher.

[0062] Optionally, embodiments of the present invention can deploy the camera in an appropriate location and then continuously transmit video and image data to a back-end system for processing. For example, the camera can be mounted on the lower side panel of a range hood.

[0063] It is understood that embodiments of the present invention may also provide a computer storage medium storing an image generation program; when the image generation program is executed by a processor, the image generation method as described in any embodiment of the present invention is implemented. The detailed process of the image generation method is as described above and will not be repeated here. Therefore, the present invention can provide an intelligent range hood equipped with a processor and a memory, wherein the memory is used to store an interference image generation program, and the processor is used to execute the interference image generation program.

[0064] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable storage medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections (electronic devices) having one or more wires, portable computer disks (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM, or flash memory), fiber optic devices, and compact disc read-only memory (CDROM). Furthermore, computer-readable storage media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0065] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0066] In the description of this specification, the references to terms such as "this embodiment," "an embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0067] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., and "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0068] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0069] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0070] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention specification and drawings under the concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. An image generation method, characterized in that, include: Acquire interference-free images, which are images of real scenes; The interference-free image is divided into multiple first sub-images by performing grid division processing. Select a second sub-image from the plurality of first sub-images, and save the remaining first sub-images; The second sub-image is subjected to interference processing to generate a third sub-image; The third sub-image is combined with the remaining first sub-image to generate an image with interference; The step of selecting a second sub-image from the plurality of first sub-images includes: Iterate through all first sub-images; The probability of interference for each first sub-image is calculated based on the probability distribution, which includes either a Gaussian distribution or a random distribution. The second sub-image is selected from all the first sub-images based on the probability of being interfered with; Determine whether the number of disturbed images generated from the current disturb-free images has reached the required iteration round; If the iteration round is not reached, the grid partitioning parameters are reselected, and the step of performing grid partitioning on the current interference-free image is returned; the grid partitioning parameters include at least one of the following: the number of rows of the grid, the number of columns of the grid, and the number of second sub-images to be determined; Alternatively, processing of the current interference-free image based on the end of the iteration round.

2. The image generation method according to claim 1, characterized in that, The step of selecting a second sub-image from the plurality of first sub-images includes: Randomly select a first sub-image as the second sub-image; Based on the connected component analysis method, a new second sub-image is selected from the first sub-images adjacent to the current second sub-image; The process can either return to the step of selecting a new second sub-image based on the current number of second sub-images, or end the second sub-image selection step.

3. The image generation method according to claim 2, characterized in that, The step of selecting a new second sub-image from the adjacent first sub-images of the current second sub-image according to the connected component analysis method includes: Get the pixel values ​​of each first sub-image adjacent to the current second sub-image; Calculate the pixel difference between each first sub-image and the current second sub-image; The first sub-image is selected as the new second sub-image based on the pixel difference.

4. The image generation method according to claim 1, characterized in that, The grid division process for the interference-free image includes: setting an odd number of rows and odd number of columns of grid in the interference-free image area, and performing equidistant grid division process on the interference-free image; The step of selecting a second sub-image from the plurality of first sub-images includes: The first sub-image located at the center of the interference-free image is taken as the second sub-image; Based on the central connected region analysis method, a new second sub-image is selected from the first sub-images adjacent to the current second sub-image; The process can either return to the step of selecting a new second sub-image based on the current number of second sub-images, or end the second sub-image selection step.

5. The image generation method according to claim 1, characterized in that, The process of performing grid division on the interference-free image includes: Construct an interference-free dataset using multiple interference-free images; Based on the non-repeating selection method, an interference-free image is obtained from the interference-free dataset for grid division processing.

6. The image generation method according to claim 1, characterized in that, The acquisition of interference-free images includes: Provides a kitchen environment with a clear view; The camera mounted on the cooking utensils is activated in the described kitchen environment; The camera captures images of the current real scene as interference-free images.

7. A method for removing interference from an image, characterized in that, include: Obtain interference-free images; Based on the interference-free image, an interference-involved image is obtained by the image generation method according to any one of claims 1 to 6; An image pair training set is formed using the disturbed image and the undisturbed image, and an adversarial generative network model is trained using the image pair training set. Use a trained generative adversarial network model to remove noise from the image to be processed.

8. An image generation apparatus, characterized in that, include: The image acquisition module is used to acquire interference-free images, which are images of real scenes; The grid division module is used to perform grid division processing on the interference-free image to divide the interference-free image into multiple first sub-images; A sub-image selection module is used to select a second sub-image from the plurality of first sub-images and save the remaining first sub-images; A sub-image interference module is used to perform interference processing on the second sub-image to generate a third sub-image; A sub-image combination module is used to combine the third sub-image with the remaining first sub-image to generate an image with interference; The sub-image selection module is also used to perform: traversing all first sub-images; calculating the probability of interference for each first sub-image according to a probability distribution, wherein the probability distribution includes either a Gaussian distribution or a random distribution; and selecting a second sub-image from all the first sub-images according to the probability of interference. The iterative judgment module is used to determine whether the number of interfering images generated from the current interference-free image has reached the required iteration round; if the iteration round has not been reached, the grid partitioning parameters are reselected, and the grid partitioning step for the current interference-free image is returned; the grid partitioning parameters include at least one of the number of rows, the number of columns, and the number of second sub-images to be determined; or, the processing of the current interference-free image ends if the iteration round has been reached.

9. A cooking utensil, characterized in that, Includes the image generation apparatus as described in claim 8.

10. A computer storage medium, characterized in that, The computer storage medium stores an image generation program; when the image generation program is executed by the processor, it implements the image generation method as described in any one of claims 1 to 6.

Images