Intelligent dish washing machine based on time sequence visual perception

Abstract

The invention relates to the technical field of dish-washing machines, in particular to an intelligent dish-washing machine based on time sequence visual perception, which comprises an image acquisition module configured to be activated in response to an extraction action of a target dish basket, continuously capturing a time sequence image sequence of the target space at a preset sampling frequency; wherein the target space is defined between the inner surface of the opened door plate and the bottom surface of the bowl basket on the last but one layer; the bowl basket below the last but one layer of bowl basket is a target bowl basket; the image acquisition module is further configured to stop capturing in response to the pushing action of the target bowl basket; the image processing module is connected with the image acquisition module and is used for receiving the time sequence image sequence, processing the time sequence image sequence by applying a preset image processing algorithm and generating a load signal representing a target space state; a control system; according to the invention, the visual shielding probability caused by tableware stacking can be obviously reduced, and accurate identification of tableware loads is realized.

Description

Technical Field

[0001] This disclosure relates to the field of dishwasher technology, and in particular to an intelligent dishwasher based on time-series visual perception. Background Technology

[0002] As a major category of modern home appliances, dishwashers have greatly reduced the burden of kitchen labor. With the continuous development of technology, people are trying to apply image recognition and algorithm models to dishwashers to improve their intelligence, ultimately achieving the goals of improving cleaning performance and user experience, as well as saving water and electricity.

[0003] For example, patent number 2023117925687 provides a dishwasher control method, device, dishwasher and storage medium. Before running the washing program, or in response to the closing of the dishwasher door, it obtains the placement information of the dish rack support inside the dishwasher, such as linear displacement information. Based on the placement information of the dish rack support, it determines whether the tableware is placed regularly, and generates a prompt message when the tableware is not placed regularly, so as to remind the user to rearrange the tableware and avoid the situation where the tableware is not placed regularly, resulting in spray dead corners and incomplete cleaning.

[0004] Patent No. 2024102350624 provides a dishwasher for processing washing appliances. It proposes to detect the characteristics of washing appliances through a washing appliance identification system, such as using visible light to identify different types and quantities of washing appliances, or using infrared to identify temperature information, and then activate a suitable washing appliance processing program to help select the optimal preset washing program for the dishwasher.

[0005] Patent No. 2024118168640 provides a method and device for controlling the operation of a dishwasher. It obtains a first phase angle image, a second phase angle image, a third phase angle image, a fourth phase angle image, and pressure data from a pressure sensor located at the bottom of the dishwasher as initial parameters. Then, it combines an oil stain detection model and a density detection model to obtain the oil stain detection results and density detection results of the tableware. These are used to optimize the cleaning parameters.

[0006] Patent No. 2025106167425 provides a dishwasher and its control method, control device, and readable storage medium. The dishwasher obtains the location of dirt on the tableware through an identification module set on the inner tub. The controller can control the first spray arm assembly to run along a preset trajectory and a preset spray angle according to the location of dirt on the tableware. The controller obtains real-time images of the tableware during the cleaning process and compares them with images recorded at the previous moment. When it is determined that the color similarity of dirt at the same location in the real-time image and the image recorded at the previous moment is higher than a preset value, the controller controls the first spray arm assembly to run to the next dirt location for cleaning.

[0007] Patent No. 2025105012554 provides a washing control method, device, dishwasher equipment, computer equipment and medium. It automatically identifies the objects placed on the dishwasher equipment by acquiring images of the placement layer and identifying the stains on the placement layer. Based on a preset image processing algorithm, it identifies the stains on the target objects in the spatial image, determines the degree of dirt on the target objects according to characteristic parameters including type, quantity and size, automatically divides the placement layer that needs to be cleaned, and controls the cleaning components to clean the areas that need to be cleaned.

[0008] However, in reality, the space in a dishwasher is very limited, and the dishes will block the view significantly when placed on it. Therefore, the above image recognition scheme cannot effectively obtain the required image and cannot achieve effective application. Summary of the Invention

[0009] To address the shortcomings or problems existing in the prior art, this disclosure provides an intelligent dishwasher based on temporal visual perception, which solves the occlusion problem in the image acquisition process of existing dishwashers.

[0010] The technical solution adopted by this disclosure to solve the above-mentioned technical problem is: an intelligent dishwasher based on temporal visual perception, comprising: The image acquisition module is configured to be activated in response to the extraction action of the target basket and to continuously capture a temporal image sequence of the target space at a preset sampling frequency; The target space is defined between the inner surface of the opened door panel and the bottom surface of the second-to-last layer of the basket; the basket below the second-to-last layer of the basket is the target basket; the image acquisition module is further configured to stop capturing in response to the pushing action of the target basket; An image processing module, connected to the image acquisition module, is used to receive the time-series image sequence and apply a predetermined image processing algorithm to process the time-series image sequence to generate a load signal characterizing the target spatial state. The control system includes an image acquisition module and an image processing module, both of which are connected to the control system. The control system receives the load signal and generates control commands based on the load signal. An action actuator, connected to the control system, is used to perform a predetermined action in response to the control command.

[0011] In a preferred embodiment, the preset sampling frequency is 10~30fps.

[0012] In a preferred embodiment, the image processing module is configured to generate a load signal through the following processing flow: image preprocessing, temporal image alignment, spatial object reconstruction, and object information fusion.

[0013] In a preferred embodiment, the image preprocessing includes performing distortion correction on the time-series image sequence and filtering keyframes based on image sharpness and motion blur.

[0014] In a preferred embodiment, the temporal image alignment includes mapping the keyframes to a unified coordinate system through feature point matching and motion estimation.

[0015] As a preferred embodiment, the spatial object reconstruction includes multi-view geometric reconstruction of aligned keyframes based on the motion recovery structure algorithm to generate a sparse three-dimensional point cloud of the target space.

[0016] As a preferred implementation, the item information fusion includes parsing the topological features of the sparse three-dimensional point cloud, matching it with a preset tableware geometric feature library, and clustering to generate a final item list.

[0017] In a preferred embodiment, the dishwasher includes a housing and a door panel hinged to the housing, and an inner liner is provided in the housing; the image acquisition module is a camera, and the optical field of view of the camera covers the entire target space without blind spots.

[0018] In a preferred embodiment, the camera is installed at an angle of θ, the axial distance from the camera to the center point of the target basket is H, and the horizontal distance from the camera to the center point of the target basket is D, all of which satisfy θ=arctan(H / D).

[0019] In some embodiments, there are two cameras, which are respectively located on two opposite side walls of the inner liner and are positioned close to the door panel; the optical field of view of the cameras satisfies the following condition: when the door panel is fully open, its field of view completely covers the target space defined by the inner surface of the door panel, the bottom surface of the second-to-last dish rack, and the edge of the dishwasher inner liner, and there is no visual obstruction area.

[0020] In some embodiments, the camera is located at the top of the dishwasher; the optical field of view of the camera satisfies the following condition: when the door is fully open, its field of view completely covers the target space defined by the inner surface of the door, the bottom surface of the second-to-last dish rack, and the edge of the dishwasher's inner liner, and there is no visual obstruction area.

[0021] As a preferred embodiment, the target bowl basket is divided into a first area, a second area, a third area and a fourth area. The four areas have equal areas and are distributed in a "field" shape. Among them, the first area and the fourth area are arranged diagonally, and the second area and the third area are arranged diagonally; the image acquisition module independently captures a sequence of timing images for each area, the image processing module processes the sequence of timing images for each area to generate a corresponding load signal, the control system receives the load signal, and generates a corresponding control instruction based on each load signal. The control instruction is to control the action execution mechanism to perform ordinary cleaning, key cleaning or efficient cleaning on the target bowl basket.

[0022] As a preferred embodiment, the loads of each area include a first threshold, a second threshold and a third threshold; Trigger conditions for efficient cleaning: There is a load in any one of the first area, the second area, the third area and the fourth area; Or there are loads in any two diagonally arranged areas; Trigger conditions for key cleaning: There are loads in any two adjacent areas, and the loads are at different thresholds; Or there are loads in more than three areas, and the load in any one area is at the third threshold; Or there are loads in more than three areas, and the loads in any two diagonally arranged areas are both at the third threshold; Trigger conditions for ordinary cleaning: There are loads in any two adjacent areas, and the loads are at the same threshold; Or there are loads in more than three areas, and the load thresholds of all areas are the same; Or there are loads in more than three areas, and the loads in any two adjacent areas are both at the third threshold.

[0023] As a preferred embodiment, the first threshold means that the number of dishes to be cleaned is 0, the second threshold means that the number of dishes to be cleaned is 1 - 3, and the third threshold means that the number of dishes to be cleaned is 4 or more.

[0024] As a preferred embodiment, the cleaning time for ordinary cleaning is t. In the ordinary cleaning mode, the spray arm assembly cleans the entire target bowl basket; the cleaning time for efficient cleaning is 50%t - t. In the efficient cleaning mode, the spray arm assembly only cleans the area where there are dishes to be cleaned; the cleaning time for key cleaning is t. In the key cleaning mode, the cleaning time for the area with more dishes to be cleaned is t1, and the cleaning time for the area with fewer dishes to be cleaned is t2, where t1 > t2 and t1 + t2 = t.

[0025] As a preferred embodiment, t1 = t2 + 20%t2.

[0026] In a preferred embodiment, the action actuator includes a drive component and a spray arm assembly, wherein the drive component is used to drive the spray arm assembly to rotate.

[0027] In a preferred embodiment, the spray arm assembly includes a main spray arm and an auxiliary spray arm connected to the main spray arm, and the drive assembly is a motor, with the drive shaft of the motor connected to the main spray arm.

[0028] Compared with existing products, this application limits the target space to the three-dimensional operating area formed when the door is fully opened (from the inner surface of the door to the bottom of the penultimate bowl basket), and uses the three-dimensional reconstruction technology of multi-frame images to trigger the acquisition of time-series images based on the pulling action of the target bowl basket, which significantly reduces the probability of visual occlusion caused by the stacking of tableware and achieves accurate identification of tableware load; relying on the ambient light introduced by the opening of the door, it ensures that the image acquisition module can effectively image the target space without the need to add a supplementary lighting module. Attached Figure Description

[0029] Figure 1 This is one of the structural schematic diagrams of an intelligent dishwasher based on temporal visual perception according to this application; Figure 2 This is the second structural schematic diagram of an intelligent dishwasher based on temporal visual perception according to this application; Figure 3 for Figure 2 A magnified view of a section at point X; Figure 4 This is the third structural schematic diagram of an intelligent dishwasher based on temporal visual perception according to this application; Figure 5 for Figure 4 A magnified view of a section at point A in the middle; Figure 6 This is a schematic diagram of the connection between the drive assembly and the main spray arm in this application; Figure 7 This is one of the algorithmic processing steps in the image processing module of this application; Figure 8 This is the second algorithmic processing step in the image processing module of this application.

[0030] In the diagram: 1. Target space; 2. Door panel; 3. Camera; 4. Main spray arm; 5. Drive assembly; 6. Drive shaft; 7. Water distribution valve assembly. Detailed Implementation

[0031] To enable those skilled in the art to better understand the technical solutions of this disclosure, the following detailed, clear, and complete description of this disclosure is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this disclosure and are not intended to limit it.

[0032] Those skilled in the art should understand that, in the disclosure of this invention, the terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not 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, the above terms should not be construed as limiting this invention.

[0033] Please refer to Figures 1-5 As shown, this application provides an intelligent dishwasher based on temporal visual perception, including an image acquisition module, an image processing module, a control system, and an action execution mechanism. The image acquisition module is configured to be activated in response to the removal action of the target basket and continuously captures a temporal image sequence of the target space 1 at a preset sampling frequency. The target space 1 is defined between the inner surface of the opened door panel 2 and the bottom surface of the second-to-last basket, and its boundary is defined by the edge of the dishwasher cabinet. The basket below the second-to-last basket is the target basket. The image acquisition module is further configured to stop capturing images in response to the pushing action of the target basket. The image processing module is connected to the image acquisition module and is used to receive the temporal image sequence and apply a predetermined image processing algorithm to process the temporal image sequence to generate a load signal characterizing the state of the target space 1. Both the image acquisition module and the image processing module are connected to the control system, which receives the load signal and generates control commands based on the load signal. The action execution mechanism is connected to the control system and is used to execute predetermined actions in response to the control commands. This application limits the target space 1 to the three-dimensional operating area (from the inner surface of the door panel 2 to the bottom of the penultimate bowl basket) formed when the door panel 2 is fully opened, and uses the three-dimensional reconstruction technology of multi-frame images to significantly reduce the probability of visual occlusion caused by the stacking of tableware, thereby achieving accurate identification of the tableware load. Relying on the ambient light source introduced by the opening of the door panel 2, the image acquisition module can effectively image the target space 1 without the need to add a supplementary lighting module.

[0034] Furthermore, the preset sampling frequency is 10~30fps. The image processing module is configured to generate the load signal through the following processing flow: image preprocessing, temporal image alignment, spatial object reconstruction, and object information fusion.

[0035] Specifically, image preprocessing includes distortion correction of the temporal image sequence and keyframe selection based on image sharpness and motion blur. Temporal image alignment involves mapping keyframes to a unified coordinate system through feature point matching and motion estimation. Spatial object reconstruction includes multi-view geometric reconstruction of the aligned keyframes based on the structure-of-motion (SOM) algorithm to generate a sparse 3D point cloud of the target space 1. Object information fusion includes parsing the topological features of the sparse 3D point cloud, matching it with a pre-defined tableware geometric feature library, and clustering to generate an object list.

[0036] like Figure 7 and Figure 8 As shown, the image processing module can generate an item list using two algorithms. Among them, Figure 7 The algorithm has high computational efficiency and is suitable for embedded devices; Figure 8 The algorithm reconstructs rich details and supports objects with complex shapes.

[0037] The dishwasher includes a shell and a door panel 2 hinged to the shell, with an inner liner inside. The image acquisition module is a camera 3, whose optical field of view covers the entire target space 1 without blind spots. The camera's installation tilt angle is θ, the axial distance from the camera to the center of the target dish rack is H, and the horizontal distance from the camera to the center of the target dish rack is D, all three satisfying θ=arctan(H / D). Specifically, θ is the angle between the optical axis of camera 3 and the horizontal plane, H is the vertical distance from the optical center of camera 3 to the center point of the target dish rack, and D is the horizontal distance from the projection point of the optical center of camera 3 on the horizontal plane to the center point of the target dish rack. This positional relationship of θ=arctan(H / D) ensures that the field of view of the two cameras 3 achieves maximum coverage within the target space 1, while avoiding image distortion and recognition blind spots caused by excessively large or small angles.

[0038] In one embodiment of this application, there are two cameras 3, which are respectively located on two opposite side walls of the inner liner and are positioned close to the door panel 2. The optical field of view of the camera 3 satisfies the following condition: when the door panel 2 is fully open, its field of view completely covers the target space 1 defined by the inner surface of the door panel 2, the bottom surface of the second-to-last dish rack, and the edge of the dishwasher inner liner, without any visual obstruction.

[0039] In another embodiment of this application, the camera 3 is a single camera located at the top of the dishwasher. The optical field of view of the camera 3 satisfies the following condition: when the door panel 2 is fully open, its field of view completely covers the target space 1 defined by the inner surface of the door panel 2, the bottom surface of the second-to-last dish rack, and the edge of the dishwasher's inner liner, without any visual obstruction. Specifically, the camera 3 is located near the door panel 2, and the camera 3 can be located at the top of the inner liner or at the top of the outer casing.

[0040] Specifically, the target dish basket is divided into a first area, a second area, a third area, and a fourth area. The four areas have equal areas and are distributed in a "field" shape. Among them, the first area and the fourth area are arranged diagonally; the second area and the third area are arranged diagonally. The image acquisition module independently captures a sequence of timing images for each area, and the image processing module processes the sequence of timing images for each area to generate corresponding load signals. The control system receives the load signals and generates corresponding control instructions based on each load signal. The control instructions are to control the action execution mechanism to perform ordinary cleaning, key cleaning, or efficient cleaning on the target dish basket. It should be noted that the first area, the second area, the third area, and the fourth area here can be achieved by means of physical partitions, such as: detachable crosses, or can also be achieved through the control system.

[0041] The loads of each area include a first threshold, a second threshold, and a third threshold. The first threshold means that the number of dishes to be cleaned is 0, the second threshold means that the number of dishes to be cleaned is 1-3, and the third threshold means that the number of dishes to be cleaned is 4 or more. Trigger conditions for efficient cleaning: There is a load in any one of the first area, the second area, the third area, and the fourth area; or there are loads in any two diagonally arranged areas; Trigger conditions for key cleaning: There are loads in any two adjacent areas, and the loads are in different thresholds; or there are loads in more than three areas, and the load in any one area is in the third threshold; or there are loads in more than three areas, and the loads in any two diagonally arranged areas are both in the third threshold; Trigger conditions for ordinary cleaning: There are loads in any two adjacent areas, and the loads are in the same threshold; or there are loads in more than three areas, and the load thresholds of all areas are the same; or there are loads in more than three areas, and the loads in any two adjacent areas are both in the third threshold. It can be understood that the image processing module processes the images collected by the image acquisition module and generates corresponding load signals, and the control system will match the preset cleaning mode according to the load signals, that is, the control system will automatically match the corresponding cleaning mode according to the load conditions of each area. For example: The load signals of the first area are A0, A1, and A2 respectively, corresponding to the first threshold, the second threshold, and the third threshold. When the number of dishes to be cleaned in the first area is 0, the load signal is A0; when the number of dishes to be cleaned in the first area is 1-3 (including 1 and 3), the load signal is A1; when the number of dishes to be cleaned in the first area is more than 4 (including 4), the load signal is A2. And so on, the load signals of the second area are B0, B1, and B2 respectively; the load signals of the third area are C0, C1, and C2 respectively; the load signals of the fourth area are D0, D1, and D2 respectively.

[0042] Specifically, the cleaning time for normal cleaning is t. In normal cleaning mode, the actuator cleans the entire target dish rack, with each area requiring 25%t of the cleaning time. The cleaning time for high-efficiency cleaning is 50%t to t, preferably 50%t to 80%t. In high-efficiency cleaning mode, the actuator only cleans the areas containing dishes to be cleaned. The cleaning time for focused cleaning is t. In focused cleaning mode, the actuator cleans areas with more dishes for t1 and areas with fewer dishes for t2, where t1 > t2 and t1 + t2 = t. Preferably, t1 = t2 + 20%t2.

[0043] like Figure 6 As shown, the actuation mechanism includes a drive assembly 5 and a spray arm assembly. The drive assembly 5 drives the spray arm assembly to rotate. Specifically, the spray arm assembly includes a main spray arm 4 and an auxiliary spray arm connected to the main spray arm 4. The drive assembly 5 is a motor, and the drive shaft 6 of the motor is connected to the main spray arm 4. The main spray arm 4 includes a spray arm section and a connecting section. The connecting section is used to connect to the water distribution valve assembly 7, and the inner cavity of the spray arm section is in communication with the inner cavity of the connecting section. The auxiliary spray arm includes at least two secondary spray heads disposed on the spray arm section, and the inner cavity of the secondary spray heads is in communication with the inner cavity of the spray arm section. The secondary spray heads are provided with several spray holes. The spray arm section may or may not have spray holes. When there are two secondary spray heads, the spray arm section is provided with spray holes. When there are three or more secondary spray heads, the spray arm section does not need to be provided with spray holes. More specifically, the drive shaft 6 of the motor is connected to the connecting section of the main spray arm 4 via a spline. The main spray arm 4 is a long, straight spray arm, while the auxiliary spray head is either a short, straight type or a three-pronged type; the specific shape of the auxiliary spray head is not limited here. To make the spray arm assembly more stable during rotation, the auxiliary spray heads are symmetrically arranged on the spray arm. The main spray arm 4 rotates under the drive of the drive assembly 5, and the auxiliary spray arm rotates together with the main spray arm 4. At the same time, the auxiliary spray arm also rotates on its own axis under the action of water pressure, that is, when the auxiliary spray arm sprays water, it will rotate relative to the main spray arm 4.

[0044] This application improves cleaning efficiency through the coupled control of space and cleaning time. Specifically, in high-efficiency cleaning mode, the control system controls the operation of drive component 5, which drives the main spray arm 4 to rotate to the loaded area. The main spray arm 4 remains below the loaded area, while the auxiliary spray arm rotates and sprays water under pressure to clean the loaded area. In intensive cleaning mode, the control system controls the operation of drive component 5, which drives the main spray arm 4 to rotate. In this mode, the cleaning time for high-load areas is longer than that for low-load areas. In normal cleaning mode, the cleaning time for each area is the same, resulting in indiscriminate cleaning of all areas. By dividing the target basket into four areas and matching different cleaning modes according to the load of each area, cleaning efficiency and cleanliness are effectively improved while reducing water waste.

[0045] For example, if both the first and fourth zones are loaded (the load thresholds can be the same or different), the spray arm assembly only cleans the first and fourth zones, with a total cleaning time of 50%t to 80%t. This is considered high-efficiency cleaning. If the first, second, and fourth zones are loaded, for example, if the number of dishes to be cleaned in the first zone is 2, in the second zone is 1, and in the fourth zone is 5, then the cleaning time for the first and fourth zones is t1, and the cleaning time for the second zone is t2. This is considered focused cleaning. It should be noted that when the spray arm assembly sprays the fourth zone, it also sprays the diagonal area of ​​the fourth zone—the first zone. Therefore, although the first zone has a smaller load, its cleaning time is the same as that of the fourth zone, which has a larger load. If the first, second, and third zones are all loaded, for example, if the number of dishes to be cleaned in the first zone is 2, in the second zone is 1, and in the third zone is 3, then the spray arm assembly cleans each zone, and the cleaning time is the same.

[0046] In one embodiment of this disclosure, a water distribution valve assembly 7 is further included. The water distribution valve assembly 7 includes a valve seat and a valve plate. The main spray arm 4 is connected to the valve seat. The valve seat is provided with a water distribution chamber. The valve plate is located at the opening of the water distribution chamber. The valve plate is connected to a driving component for driving its rotation. The driving component is electrically connected to the control system. Specifically, the driving component is a motor. The output shaft of the motor is connected to the valve plate, driving the valve plate to rotate. It is understood that the dishwasher is also provided with an inner water pipe, a middle spray arm, and an upper spray arm. The middle spray arm and the upper spray arm are both connected to the inner water pipe. The middle spray arm and the upper spray arm are used to clean the dish rack in non-target areas. The inner water pipe and the main spray arm 4 are both connected to the water distribution chamber. The driving component controls whether water flows through the inner water pipe and / or the main spray arm 4 by controlling the rotation of the valve plate. The connection methods of the upper spray arm, the middle spray arm, and the inner water pipe, as well as the connection between the inner water pipe, the main spray arm 4, and the valve seat, are all prior art and will not be described in detail here.

[0047] The present application has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present application. The descriptions of the embodiments above are only for the purpose of helping to understand the present application and its core ideas. It should be noted that those skilled in the art can make several improvements and modifications to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the protection scope of the claims of the present application.

Claims

1. A smart dishwasher based on temporal visual perception, characterized in that, Comprising: An image acquisition module, configured to be activated in response to the extraction action of the target bowl basket, and continuously capture a sequence of time-series images of the target space (1) at a preset sampling frequency; Wherein, the target space (1) is defined between the inner surface of the opened door panel (2) and the bottom surface of the second-to-last bowl basket; the bowl basket below the second-to-last bowl basket is the target bowl basket; the image acquisition module is further configured to stop capturing in response to the pushing action of the target bowl basket; An image processing module, connected to the image acquisition module, for receiving the sequence of time-series images, and applying a predetermined image processing algorithm to process the sequence of time-series images to generate a load signal characterizing the state of the target space (1); A control system, both the image acquisition module and the image processing module are connected to the control system, the control system receives the load signal, and generates a control instruction based on the load signal; An action execution mechanism, connected to the control system, for executing a predetermined action in response to the control instruction.

2. The intelligent dishwasher based on temporal visual perception according to claim 1, characterized in that, The preset sampling frequency is 10~30fps.

3. The intelligent dishwasher based on temporal visual perception according to claim 1, characterized in that, The image processing module is configured to generate a load signal through the following processing flow: image preprocessing, time-series image alignment, spatial object reconstruction, and object information fusion.

4. The intelligent dishwasher based on temporal visual perception according to claim 1, characterized in that, The dishwasher includes a housing and a door panel (2) hinged to the housing, and an inner liner is provided in the housing; the image acquisition module is a camera (3), and the optical field of view of the camera (3) covers the entire target space (1) without blind spots.

5. The intelligent dishwasher based on temporal visual perception according to claim 8, characterized in that, The installation inclination angle of the camera (3) is θ, the axial distance from the camera (3) to the center point of the target bowl basket is H, and the horizontal distance from the camera (3) to the center point of the target bowl basket is D, and the three satisfy θ = arctan(H / D).

6. The intelligent dishwasher based on temporal visual perception according to claim 5, characterized in that, There are two cameras (3), and the two cameras (3) are respectively arranged on two opposite side walls in the inner liner, and the camera (3) is arranged close to the door panel (2); the optical field of view of the camera (3) satisfies: when the door panel (2) is in the fully opened state, its field of view completely covers the target space (1) defined by the inner surface of the door panel (2), the bottom surface of the second-to-last bowl basket, and the edge of the dishwasher inner liner.

7. The intelligent dishwasher based on temporal visual perception according to claim 5, characterized in that, The camera (3) is arranged at the top of the dishwasher; the optical field of view of the camera (3) satisfies: when the door panel (2) is in the fully opened state, its field of view completely covers the target space (1) defined by the inner surface of the door panel (2), the bottom surface of the second-to-last bowl basket, and the edge of the dishwasher inner liner.

8. The intelligent dishwasher based on temporal visual perception according to claim 4, characterized in that, The target bowl basket is divided into a first area, a second area, a third area, and a fourth area. The areas of the four areas are equal and are distributed in a "field" shape. Among them, the first area and the fourth area are arranged diagonally; the second area and the third area are arranged diagonally; the image acquisition module independently captures a sequence of time-series images for each area, the image processing module processes the sequence of time-series images for each area to generate a corresponding load signal, the control system receives the load signal, and generates a corresponding control instruction based on each load signal, and the control instruction is to control the action execution mechanism to perform ordinary cleaning or key cleaning or efficient cleaning on the target bowl basket; 9. The intelligent dishwasher based on temporal visual perception according to claim 8, characterized in that, The loads of each area include a first threshold, a second threshold, and a third threshold; Triggering conditions for efficient cleaning: There is a load in any of the first, second, third, and fourth regions; Or any two areas arranged diagonally may have a load; Triggering conditions for focused cleaning: There is a load in any two adjacent regions, and the load is at different thresholds; Or there is load in more than three regions, and the load in any region is at the third threshold; Or more than three areas have loads, and the loads of any two diagonally arranged areas are at the third threshold; Triggering conditions for normal cleaning: There is a load in any two adjacent regions, and the load is at the same threshold. Or there is load in more than three regions, and all regions have the same load threshold; Or there is load in more than three regions, and the load in any two adjacent regions is at the third threshold.

10. The intelligent dishwasher based on temporal visual perception according to claim 9, characterized in that, The first threshold refers to a quantity of 0 tableware to be cleaned, the second threshold refers to a quantity of 1 to 3 tableware to be cleaned, and the third threshold refers to a quantity of 4 or more tableware to be cleaned.

11. The intelligent dishwasher based on temporal visual perception according to claim 9, characterized in that, The cleaning time for the normal cleaning is t. In the normal cleaning mode, the spray arm assembly cleans the entire target dish rack. The cleaning time for the high-efficiency cleaning is 50%t~t. In the high-efficiency cleaning mode, the spray arm assembly only cleans the area with dishes to be cleaned. The cleaning time for the focused cleaning is t. In the focused cleaning mode, the cleaning time for the area with more dishes to be cleaned is t1, and the cleaning time for the area with fewer dishes to be cleaned is t2, where t1>t2, and t1+t2=t.

12. The intelligent dishwasher based on temporal visual perception according to claim 11, characterized in that, t1 = t2 + 20%t2.

13. The intelligent dishwasher based on temporal visual perception according to claim 1, characterized in that, The action execution mechanism includes a drive component (5) and a spray arm assembly, wherein the drive component (5) is used to drive the spray arm assembly to rotate.

Images