Garbage sorting robot based on deep learning

Abstract

The invention discloses a garbage sorting robot based on deep learning. The garbage sorting robot comprises a rack main body, an advancing mechanism, a mechanical arm mechanism, a garbage storage mechanism and a control host. According to the invention, complex image data in the garbage heap can be captured and analyzed in real time, the images are rapidly converted into accurate classification information by means of the strong power of deep learning, and recyclable paper, plastics, metals, harmful batteries, chemicals and other types of garbage can be accurately distinguished. By means of the process, the classification accuracy is greatly improved, the processing speed is remarkably increased, and the situation that traditional manual sorting is low in efficiency and prone to errors is thoroughly changed.

Description

Technical Field

[0001] The present invention relates to a robot, in particular to a garbage sorting robot based on deep learning, belonging to the technical field of garbage sorting. Background Art

[0002] With the rapid development of modern urban science and technology and the continuous improvement of residents' living standards, the amount of domestic garbage has shown an explosive growth trend. The randomly discarded garbage has not only seriously damaged the clean appearance and ecological environment of the city, but also posed an unprecedented challenge to the traditional garbage sorting and treatment methods that rely on manual labor. Although the traditional mode of manual sorting has played an important role in the past, facing the huge amount of garbage disposal today, its inherent disadvantages such as time-consuming, labor-intensive and inefficient are becoming more and more prominent, and it is difficult to effectively cope with large-scale and high-intensity garbage disposal needs. At the same time, with the development of social economy, labor costs continue to rise, making the economic cost of manual sorting continue to increase, bringing a heavy economic burden to the entire garbage disposal industry. More importantly, human factors such as fatigue caused by long-term work and difficulty in continuous concentration can easily cause errors in the sorting process, thereby affecting the efficiency and final treatment effect of subsequent garbage disposal, especially in the face of garbage with complex and varied compositions, a wide variety and different characteristics. The limitations of manual sorting are even more exposed, and the classification results are often difficult to achieve ideal conditions; in addition, the shortcomings of traditional garbage disposal methods in terms of environmental protection cannot be ignored. Due to incomplete sorting or incorrect classification, a large number of recyclable resources are mistakenly mixed with other garbage, which not only causes a huge waste of resources, but may also be polluted due to improper handling; at the same time, if harmful substances are mishandled, it is more likely to cause serious environmental pollution problems and cause long-term and irreversible damage to the ecological environment. In addition, potential safety hazards in the manual sorting process, such as scratches from sharp objects and direct contact with toxic and harmful substances, also pose a serious threat to the health of people engaged in sorting work and increase occupational risks.

[0003] In view of the many drawbacks of manual picking and sorting, the automation and intelligent transformation in the field of mobile garbage picking has become an inevitable trend in the development of the industry.

[0004] In recent years, garbage sorting robot technology has made significant progress. These robots can accurately identify and sort out various types of garbage, including recyclables, hazardous waste, wet garbage, etc. by integrating advanced technologies such as advanced sensors, machine vision, and artificial intelligence algorithms. Compared with traditional manual sorting, garbage sorting robots have higher sorting efficiency, lower error rate, and stronger environmental adaptability, which effectively improves the automation and intelligence level of garbage disposal. In the future, garbage sorting robot technology will continue to develop in the direction of intelligence and precision. With the continuous maturity of artificial intelligence technologies such as machine learning and deep learning, garbage sorting robots will have stronger autonomous learning capabilities, and will be able to continuously optimize sorting algorithms and improve sorting accuracy and efficiency. At the same time, combined with technologies such as the Internet of Things and big data, garbage sorting robots will be interconnected with other devices to build a more complete garbage disposal ecosystem.

[0005] The garbage sorting mobile robot is an innovative robot platform that integrates a robot chassis, a control system module, and an intelligent perception system module. Its design process includes three aspects: mechanics, control, and intelligent perception.

[0006] First, in terms of mechanics, the use of a wheeled chassis as a mobile platform enables the robot to move freely on various surfaces and in various environments. The wheeled chassis ensures that the robot has good passability and stability through the combination of tires, drive devices and suspension systems. Secondly, the actuator is a key component of the robot and is responsible for completing various tasks. It can be a flexible robotic arm for fine movements such as handling, assembly and operation; it can also be other types of actuators, such as various tools and sensors, to meet the needs of different tasks. The design and control capabilities of the actuator directly affect the function and performance of the robot. Finally, in terms of control, the control system plays the role of the "brain" of the robot, responsible for real-time monitoring and control of the robot's movement and operation. This requires combining the research content of the three aspects of mechanics, control and intelligent perception to ensure the rationality of the robot system design, and complete the processing and installation of the mechanical parts of the robot system and the overall motion debugging.

[0007] In the prior art, the intelligent garbage sorting robot disclosed in announcement number CN116985090A includes a motion module, an action module and a control system; the motion module includes a chassis walking device consisting of four Mecanum wheels and a motor; the action module includes a six-axis robotic arm and a robotic claw; the wrist of the six-axis robotic arm is equipped with an Intel RealSense D455 binocular camera; the bottom control system completes the control of the motor and the servo and the obstacle avoidance sensor signal processing, and the upper processing system completes SLAM mapping, positioning and navigation, garbage classification, positioning and path planning of the robotic arm grasping based on laser radar and ROS, and the information processed by the upper processing system is sent to the bottom control system through serial port communication to adjust the speed of the motor and control the various servos of the robotic arm, thereby controlling the movement of the robot and the movement of the robotic arm. The present invention identifies the domestic garbage in the designated area, distinguishes the type of garbage it belongs to, completes the positioning, picking and classification operations, and realizes the fully automatic unmanned garbage sorting work in the designated area through the mobile search of the robot. Although some automated sorting equipment has been put into use in the current market, its popularity is still insufficient, and the technical level is uneven, making it difficult to fully cover and meet the urgent needs of the garbage disposal industry for efficient and accurate sorting. What is particularly critical is that the lack of support from advanced intelligent recognition systems and efficient algorithms makes it difficult for existing systems to quickly and accurately identify and sort out various types of garbage, especially those with different shapes and complex characteristics, thereby limiting the further improvement of garbage disposal efficiency and accuracy. In addition, although existing robots can realize automatic grasping, transportation and sorting of garbage, the sorting efficiency is low, and the sorting accuracy is also low, which cannot actually and effectively alleviate the pressure of manual sorting. The garbage classification image feature database cannot be improved, and the sorting strategy cannot be optimized using big data analysis technology, which leads to the inability to improve the adaptability and stability of the system, and it is impossible to ensure that efficient and accurate sorting can be achieved in different scenarios. Summary of the invention

[0008] The purpose of the present invention is to provide a garbage sorting robot based on deep learning in order to solve at least one of the above-mentioned technical problems.

[0009] The present invention achieves the above-mentioned purpose through the following technical solutions: A garbage sorting robot based on deep learning, comprising a frame body, a traveling mechanism, a mechanical arm mechanism, a garbage storage mechanism and a control host, wherein the traveling mechanism is arranged at the lower part of the frame body, the mechanical arm mechanism is arranged at the front end of the frame body, the garbage storage mechanism is arranged on both sides of the frame body, and the storage mechanism is connected to the frame body, and the control host is installed on the frame body; A garbage identification system is integrated in the control host. The garbage identification system consists of a navigation module, a target identification module, a sorting control module, a main control chip, a wireless communication module and a host computer module. The navigation module, the target identification module, the sorting control module, the wireless communication module and the host computer module are all connected to the main control chip for signal transmission.

[0010] As a further solution of the present invention: the frame body includes two supporting bottom plates, the docking bottom surfaces of the two supporting bottom plates are connected to a docking seat, the robotic arm mechanism and the garbage storage mechanism are installed on the upper plate surface of one of the supporting bottom plates of the frame body, the control host is installed on the upper plate surface of the other supporting bottom plate of the frame body, and the lower plate surface of the other supporting bottom plate is fixedly connected to a battery.

[0011] As a further solution of the present invention: the travel mechanism includes a reduction motor and a Mecanum wheel, the Mecanum wheels are rotatably connected to the four corners of the frame body, and the four Mecanum wheels are symmetrically arranged with each other, each Mecanum wheel is matched with a reduction motor, and the reduction motor and the Mecanum wheel are connected by a coupling.

[0012] As a further solution of the present invention: the mechanical arm structure includes a mechanical claw, a servo, a pan-tilt head and a connecting rod, both ends of the connecting rod are respectively connected to the servo, three servos are provided, the pan-tilt head is fixedly connected to the frame body, the servo at the bottom end is installed on the pan-tilt head, and the servo at the top end is connected to the mechanical claw.

[0013] As a further solution of the present invention: two garbage storage mechanisms are provided, the garbage storage mechanisms include a main box body, a detachable garbage basket and a hinged door, the upper end of the main box body has no cover, the detachable garbage basket is built into the main box body, and the hinged door is movably connected to one side of the main box body.

[0014] As a further solution of the present invention: the target recognition module of the garbage identification system performs target detection and target classification on the images obtained by the camera through the SSD_Mobilenet_V2 deep learning algorithm. The sorting control module of the garbage identification system includes a camera and a contact sensor. The camera is installed directly in front of the uppermost servo, and the contact sensor is connected to the center of the mechanical claw.

[0015] As a further solution of the present invention: the navigation module of the garbage identification system includes a laser radar, and the laser radar is installed at the bottom of the gimbal.

[0016] The beneficial effect of the present invention is that the present invention can capture and analyze complex image data in garbage dumps in real time. With the powerful power of deep learning, these images are quickly converted into accurate classification information, whether it is recyclable paper, plastic, metal, or harmful batteries, chemicals, or other types of garbage, they can be accurately distinguished. This process not only greatly improves the accuracy of classification, but also significantly speeds up the processing speed, completely changing the inefficiency and error-prone situation of traditional manual sorting.

[0017] The present invention has the ability to perceive the environment through high-precision laser radar, can autonomously plan paths, flexibly avoid various obstacles, and ensure the safe and efficient operation of the platform. Combined with the dual protection of GPS and inertial navigation systems, the platform can achieve accurate positioning and path optimization even in indoor or complex terrain without GPS signals, greatly expanding the boundaries of its application scenarios.

[0018] The present invention achieves real-time uploading of working status and sorting data to the cloud server. This innovative measure enables managers to transcend the limitations of time and space, remotely monitor the operation of the platform through mobile phone APP or computer, and adjust work strategies and optimize resource allocation based on real-time data. At the same time, the powerful data analysis function also provides a scientific basis for urban waste management, helping managers make more intelligent decisions.

[0019] The present invention adopts a modular structure, which gives it extremely high flexibility and scalability. According to actual needs, the platform can easily configure modules with different functions, such as adding a robotic arm to achieve more sophisticated grasping operations, or adding sensors to enhance environmental perception capabilities. In addition, with the continuous optimization of machine learning algorithms and the continuous emergence of new technologies, the platform can also be easily upgraded to always maintain technological leadership and meet more diversified garbage disposal needs in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the overall front view structure of the present invention; Figure 2 It is a schematic diagram of the bottom structure of the present invention; Figure 3 It is a schematic diagram of the overall rear view structure of the present invention.

[0021] In the figure: 1. frame body, 2. mechanical claw, 3. camera, 4. main box, 5. servo, 6. control host, 7. detachable trash basket, 8. hinge door, 9. reduction motor, 10. Mecanum wheel, 11. pan / tilt, 12. battery, 13. docking station, 14. laser radar, 15. contact sensor, 16. connecting rod. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Embodiment 1, as Figures 1 to 3 As shown, a garbage sorting robot based on deep learning includes a frame body 1, a traveling mechanism, a mechanical arm mechanism, a garbage storage mechanism and a control host 6, wherein the traveling mechanism is arranged at the lower part of the frame body 1, the mechanical arm mechanism is arranged at the front end of the frame body 1, the garbage storage mechanism is arranged on both sides of the frame body 1, and the storage mechanism is connected to the frame body 1, the control host 6 is installed on the frame body 1, and the control command issued by the control host 6 can enable the traveling mechanism to drive the robot to walk, and the mechanical arm mechanism is used for clamping and lifting domestic garbage and transporting it to the garbage storage mechanism, and the garbage is clamped, lifted and rotated by the mechanical arm structure and sent into the garbage storage mechanism; A garbage identification system is integrated in the control host 6. The garbage identification system consists of a navigation module, a target identification module, a sorting control module, a main control chip, a wireless communication module and a host computer module. The navigation module, the target identification module, the sorting control module, the wireless communication module and the host computer module are all connected to the main control chip by signal transmission. The control host 6 plays the role of the "brain" of the robot, responsible for real-time monitoring and controlling the movement and operation of the robot, ensuring the rationality of the robot system design, and completing the processing and installation of mechanical parts of the robot system and the overall movement, realizing an automated and intelligent garbage sorting process, which not only greatly reduces the labor intensity of sanitation workers, but also significantly improves the resource recovery rate and reduces the environmental pollution caused by landfill and incineration of garbage.

[0024] Embodiment 2, in addition to all the technical features of Embodiment 1, this embodiment also includes: the frame body 1 includes two supporting bottom plates, the docking bottom surfaces of the two supporting bottom plates are connected to a docking seat 13, the robot arm mechanism and the garbage storage mechanism are installed on the upper plate surface of one of the supporting bottom plates of the frame body 1, the control host 6 is installed on the upper plate surface of the other supporting bottom plate of the frame body 1, and the lower plate surface of the other supporting bottom plate is fixedly connected to a battery 12, so that the frame body 1 can be disassembled through the docking seat 13, thereby facilitating the assembly of the robot.

[0025] The traveling mechanism includes a reduction motor 9 and a Mecanum wheel 10. The Mecanum wheels 10 are rotatably connected to the four corners of the frame body 1, and the four Mecanum wheels 10 are symmetrically arranged. Each Mecanum wheel 10 is matched with a reduction motor 9. The reduction motor 9 and the Mecanum wheel 10 are connected through a coupling. The reduction motor 9 drives the Mecanum wheel 10 to move, so that the robot has good maneuverability and flexibility, and adopts a four-wheel drive mode to adapt to various road surfaces and various paths.

[0026] The mechanical arm structure includes a mechanical claw 2, a servo 5, a pan-tilt 11 and a connecting rod 16. Both ends of the connecting rod 16 are respectively connected to the servo 5. Three servos 5 are provided. The pan-tilt 11 is fixedly connected to the frame body 1. The servo 5 at the bottom is installed on the pan-tilt 11. The servo 5 at the top is connected to the mechanical claw 2. The plurality of servos 5 and the plurality of connecting rods 16 are interconnected, which are used for steering the mechanical arm and lifting the mechanical claw 2, and the garbage is grabbed by the clamping of the mechanical claw 2.

[0027] There are two garbage storage mechanisms, which include a main box body 4, a detachable garbage basket 7 and a hinge door 8. The upper end of the main box body 4 has no cover, and the detachable garbage basket 7 is built in the main box body 4. The hinge door 8 is movably connected to one side of the main box body 4. The position of the detachable garbage basket 7 is fixed by the main box body 4, and the detachable garbage basket 7 can be easily taken out and placed by opening the hinge door 8, so that the garbage placed in the garbage basket can be easily cleaned.

[0028] Embodiment 3. In addition to all the technical features of Embodiment 1, this embodiment also includes: the target recognition module of the garbage identification system performs target detection and target classification on the image obtained by the camera through the SSD_Mobilenet_V2 deep learning algorithm. The sorting control module of the garbage identification system includes a camera 3 and a contact sensor 15. The camera 3 is installed in front of the uppermost servo 5. The contact sensor 15 is connected to the center of the claw of the mechanical claw 2 so that the garbage to be grabbed can be photographed by the camera 3. The image obtained by the camera 3 is subjected to target detection and target classification through the SSD_Mobilenet_V2 deep learning algorithm. The coordinates and angle information of the target are obtained as input information of the sorting control module, and then the mechanical claw 2 is controlled to aim at the garbage. When the contact sensor 15 touches the garbage, it means that the mechanical claw 2 moves into place. At this time, it can be ensured that the mechanical claw 2 grabs the garbage more firmly, and the sorting control module can execute the placement of the grabbed garbage into the corresponding garbage storage mechanism.

[0029] The navigation module of the garbage identification system includes a laser radar 14, which is installed at the bottom of the gimbal 11. The laser radar 14 is used to collect environmental information of the cleaning area, realize the SLAM function based on the scan matching algorithm, and traverse the cleaning area through the optimal path algorithm for path planning. The laser radar 14 uses an 8000 sampling frequency to ensure the quality of map construction when the robot moves at a high speed. The ranging core of the Silan radar A1 rotates clockwise to achieve 360° all-round scanning detection. Then, the contour map of the surrounding environment is obtained and a 2D plane image is constructed. In order to enable the robot to navigate at the fastest speed, the robot uses the latest Raspberry Pi 5 4B board to process navigation, which is light, easy to carry and has good performance. The single-chip microcomputer uses STM32F405, which is mainly used to control the deceleration DC motor. Using a high-frequency (168MHz) STM32F4 can increase the reaction speed and make walking more coordinated.

[0030] Embodiment 4. This embodiment provides an identification and control method for a garbage sorting robot based on deep learning. The garbage sorting robot mentioned in this embodiment is the garbage sorting robot based on deep learning mentioned in Embodiments 1 to 3. When in use, ensure that the garbage sorting robot is in patrol mode and is placed on a road surface in a spacious area with a suitable environment (community, park, etc.).

[0031] Furthermore, the automatic mobile garbage sorting robot is placed at a designated starting position, ensuring that it has unimpeded access to a power source or that the battery is sufficiently charged, and the robot is started through a remote control, a mobile phone APP or a preset program to perform self-inspection and calibration, including initialization of sensors, cameras, navigation systems, etc.

[0032] Furthermore, the robot uses built-in laser radar, camera and SLAM (simultaneous localization and mapping) technology to autonomously navigate in the venue, autonomously plan paths, avoid obstacles, and ensure safe and efficient movement. It also captures garbage images through high-definition cameras and uses deep learning algorithms to quickly analyze the images and identify the types of garbage. This process relies on a pre-trained model that can accurately distinguish different types of garbage.

[0033] Furthermore, once the visual module identifies the type of garbage, the robot will use the integrated mechanical arm or gripping device to accurately grab or absorb the target garbage based on the identification result. At the same time, according to the identified type of garbage, the robot will place the garbage into the corresponding classified garbage bin. This process may involve the robot dynamically adjusting its position and direction to ensure the accuracy of placement.

[0034] Furthermore, during the entire sorting process, the robot will record sorting data in real time, including sorting speed, accuracy, quantity of various types of garbage, etc., to provide data support for subsequent operational optimization. Through the cloud platform or local control system, users can view the robot's working status and sorting report, and adjust parameters or upgrade software as needed to continuously improve sorting efficiency and accuracy.

[0035] Furthermore, regularly inspect and maintain the robot's mechanical parts, sensors, batteries, etc. to ensure they are in good working condition. Clean the dirt and garbage residue on the robot's surface, and keep the camera and sensors clean to avoid affecting recognition and navigation performance.

[0036] Furthermore, when the power level is below the set threshold, the robot will automatically return to the charging station for charging without human intervention. When the robot is not needed, it can be manually turned off through the remote control or APP to save energy. When the device starts the automatic patrol mode, camera 3, as the core component of the visual system, instantly captures environmental images and autonomously analyzes and identifies garbage targets within the field of view.

[0037] Once the target garbage is successfully identified, the device immediately starts the garbage sorting process. The process first uses the camera 3 to accurately aim at the target garbage, collect its image information in real time, and upload it to the cloud server immediately. The automatic recognition mechanism embedded in the cloud server quickly processes these image data and uses deep learning algorithms for efficient analysis to determine the specific type of garbage.

[0038] Then, based on the instructions returned by the cloud, the control host 6 built into the device immediately plans an optimal path to guide the mechanical body to approach the target garbage smoothly and efficiently. During the approach process, the target recognition module accurately extracts the contour information of the target garbage, and sets multiple sampling points on this basis to construct a detailed set of garbage feature points. By mathematically fitting these feature points, the device can generate an accurate picking trajectory.

[0039] Then, the robot arm mechanism accurately completes the picking and lifting operation of the target garbage according to the picking trajectory. This process not only demonstrates the high-precision control capability of the device, but also demonstrates its flexible response capability in complex environments.

[0040] After successful collection, the device will perform preliminary classification of the garbage based on the results of big data analysis, and automatically place it in a pre-set garbage storage mechanism that matches the type of garbage. This process not only improves the accuracy of garbage classification, but also greatly improves the overall work efficiency and automation level.

[0041] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0042] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A garbage sorting robot based on deep learning, comprising a frame body (1), a travel mechanism, a mechanical arm mechanism, a garbage storage mechanism and a control host (6), characterized in that: The travel mechanism is arranged at the lower part of the frame body (1), the mechanical arm mechanism is arranged at the front end of the frame body (1), the garbage storage mechanism is placed on both sides of the frame body (1), and the storage mechanism is connected to the frame body (1), and the control host (6) is installed on the frame body (1); The control host (6) is integrated with a garbage identification system, which is composed of a navigation module, a target identification module, a sorting control module, a main control chip, a wireless communication module and a host computer module. The navigation module, the target identification module, the sorting control module, the wireless communication module and the host computer module are all connected to the main control chip for signal transmission.

2. The garbage sorting robot according to claim 1, characterized in that: The frame body (1) comprises two supporting bottom plates, the butt bottom surfaces of the two supporting bottom plates are connected to a butt socket (13), the mechanical arm mechanism and the garbage storage mechanism are mounted on the upper plate surface of one of the supporting bottom plates of the frame body (1), the control host (6) is mounted on the upper plate surface of the other supporting bottom plate of the frame body (1), and the lower plate surface of the other supporting bottom plate is fixedly connected to a storage battery (12).

3. The garbage sorting robot according to claim 1, characterized in that: The travel mechanism comprises a reduction motor (9) and a Mecanum wheel (10), wherein the Mecanum wheels (10) are rotatably connected to the four corners of the frame body (1), and the four Mecanum wheels (10) are symmetrically arranged with respect to each other, each Mecanum wheel (10) is matched with a reduction motor (9), and the reduction motor (9) and the Mecanum wheel (10) are connected via a coupling.

4. The garbage sorting robot according to claim 1, characterized in that: The mechanical arm structure comprises a mechanical claw (2), a steering gear (5), a pan-tilt platform (11) and a connecting rod (16), wherein both ends of the connecting rod (16) are respectively connected to the steering gear (5), three steering gears (5) are provided, the pan-tilt platform (11) is fixedly connected to a frame body (1), the steering gear (5) at the bottom end is mounted on the pan-tilt platform (11), and the steering gear (5) at the top end is connected to the mechanical claw (2).

5. The garbage sorting robot according to claim 1, characterized in that: The garbage storage mechanism is provided with two, and the garbage storage mechanism comprises a main box body (4), a detachable garbage basket (7) and a hinged door (8); the upper end of the main box body (4) has no cover, the detachable garbage basket (7) is built in the main box body (4), and the hinged door (8) is movably connected to one side of the main box body (4).

6. The garbage sorting robot according to claim 4, characterized in that: The target recognition module of the garbage recognition system performs target detection and target classification on the image acquired by the camera through the SSD_Mobilenet_V2 deep learning algorithm. The sorting control module of the garbage recognition system comprises a camera (3) and a contact sensor (15). The camera (3) is installed in front of the uppermost steering gear (5). The contact sensor (15) is connected to the center of the mechanical claw (2).

7. The garbage sorting robot according to claim 6, characterized in that: The navigation module of the garbage identification system comprises a laser radar (14), and the laser radar (14) is installed at the bottom end of the pan-tilt platform (11).

Images