Vehicle-mounted camera self-cleaning device and method based on ultrasonic vibration and scene classification

Abstract

The invention discloses a vehicle-mounted camera self-cleaning device and method based on ultrasonic vibration and scene classification. The vehicle-mounted camera self-cleaning device comprises a lens, a camera assembly, a piezoelectric transducer, a camera shell, a PCB, a heating coil, a front shell and a rear shell. The front shell and the rear shell are located at the front end and the rear end of the camera shell respectively and fixedly connected, the lens is fixed between the front shell and the camera shell, the piezoelectric transducer is fixed between the lower side of the lens and a groove formed in the camera shell, the heating coil is installed on the inner side of the lens, and the camera assembly is located in the camera shell. A PCB is fixedly installed at the lower end of the camera assembly, the temperature of the whole lens is increased through a heating coil surrounding the lens, and meanwhile pollutants are shaken off from the range of the lens through rapid ultrasonic vibration through a piezoelectric ceramic piece surrounding and bonded to the lower portion of the lens. The device has the advantages of being simple in structure, convenient to assemble, good in reliability, small in size and low in cost.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle-mounted equipment, and in particular to a vehicle-mounted camera self-cleaning device and method based on ultrasonic vibration and scene classification. Background Art

[0002] In a known product (patent number: CN202010646602), a heating device is set inside the lens, located at the edge of the lens body, and is used to transfer heat to the lens body after power is supplied, wherein the internal structure of the heating device includes heating elements, which are embedded in a ceramic heating ring or a polyimide (PI) heating film. After the external power supply is supplied, the heating device generates heat and directly transfers the heat to the lens, thereby realizing the rain and snow removal function.

[0003] In a known product (patent number: CN202110788523), the heating element includes an electrode and a conductive heating component. The conductive heating component is placed in the edge area of ​​the lens and is tightly attached to the surface of the lens by evaporation or sputtering technology. The connecting end of the electrode is electrically connected to the conductive heating component, and the power connection end is connected to a power source to provide electrical energy to the conductive heating component. When powered on, the conductive heating component generates heat and transfers the heat to the lens, thereby realizing the rain and snow removal function.

[0004] Identified shortcomings of the prior art: 1) Placing the heating device inside the lens requires a certain amount of space to be reserved during design, which makes it difficult to miniaturize the lens and meet the growing demand for compactness. At the same time, the use of an external power supply requires additional consideration of the lead-out of the wires, which not only increases the complexity of manufacturing and assembly, but also increases the overall cost accordingly.

[0005] 2) Removal by gas, liquid, heating and other methods requires the installation of additional nozzles, cleaning pumps, solenoid valves, water tanks and other equipment. The installation cost is high, the structure is complex, and it cannot cope with pollutant coverage other than rain and snow. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention discloses a vehicle-mounted camera self-cleaning device and method based on ultrasonic vibration and scene classification to solve the problems raised in the above background technology.

[0007] To achieve the above-mentioned objectives, the present invention provides the following technical solutions: a self-cleaning device for a vehicle-mounted camera based on ultrasonic vibration and scene classification, comprising a lens, a camera assembly, a piezoelectric transducer, a camera housing, a PCB circuit board, a heating coil, a front shell and a rear shell; the front shell and the rear shell are respectively located at the front and rear ends of the camera housing and are fixedly connected, and a accommodating cavity is formed therein after the camera housing and the rear shell are fixedly connected, and the lens is fixed between the front shell and the camera housing, so that the vibration generated by the piezoelectric transducer can be evenly transmitted to various parts of the lens, thereby ensuring comprehensive cleaning of the lens. The piezoelectric transducer is fixed between the lower side of the lens and the groove in the camera housing. A piezoelectric ceramic sheet is arranged inside the piezoelectric transducer. The heating coil is installed on the inner side of the lens. The camera assembly is located at the center position inside the camera housing to ensure its stability during vibration and heating and will not affect the cleaning operation of the lens. A PCB circuit board is fixedly installed at the lower end of the camera assembly. The lower side of the PCB circuit board is fixedly connected to the rear shell. A plug connector for electrically connecting to the vehicle body is fixedly connected to the rear shell. The plug connector is electrically connected to the PCB circuit board for powering the camera or transmitting information. A connecting wire electrically connected to the piezoelectric transducer and the heating coil is fixedly connected to the camera housing. The other end of the connecting wire is electrically connected to the PCB circuit board. The entire lens is heated by the heating coil surrounding the lens. At the same time, pollutants are shaken out of the lens range by surrounding the piezoelectric ceramic sheet bonded under the lens.

[0008] Preferably, the piezoelectric transducer is directly fixed to the lower side of the lens by bonding, hot pressing or welding. When the working environment temperature is lower than 50°C and the humidity is lower than 80%, bonding is preferably used to connect the piezoelectric transducer and the lens, which is simple to operate and low in cost. When the temperature is higher than 50°C, hot pressing is used to ensure the stability of the connection. When higher connection strength and electrical performance are required, such as in a high vibration environment, welding is used. The heating coil is installed on the inner side of the lens and connected to the PCB circuit board by bonding, hot pressing or welding. One end of the piezoelectric transducer uses a sticky material with thermal conductivity. After the piezoelectric transducer is wrapped around the camera housing, the two ends are bonded or hot pressed. For the heating coil, when quick installation is required and the electrical performance requirements are not particularly high, bonding is used. When higher reliability and long-term stability are required, hot pressing or welding is used.

[0009] Preferably, mechanical vibration is generated by a piezoelectric ceramic sheet, and the material of the piezoelectric ceramic sheet includes potassium sodium niobate or barium strontium niobate material. The potassium sodium niobate material has a higher piezoelectric constant and can generate larger mechanical vibrations at a lower voltage, and is suitable for cleaning scenarios with higher requirements on vibration intensity; the barium strontium niobate material has good temperature stability and can maintain stable piezoelectric properties within a wide temperature range, and is suitable for vehicle-mounted camera self-cleaning devices operating at different ambient temperatures.

[0010] Preferably, one side of the piezoelectric ceramic sheet is integrated with or uses a sticky material with thermal conductivity to fix the piezoelectric ceramic sheet and the lens. When the sticky material is sticky at room temperature, direct bonding is performed. When the sticky material needs to be melted at high temperature to become sticky, hot pressing is used to fix the lens and the heating element.

[0011] Preferably, the heating coil obtains power and generates heat through the PCB circuit board, causing the lens to heat up. The heating coil is arranged around the inner ring of the lens barrel of the vehicle-mounted lens, and the heating coil is in the form of a flexible circuit board or a ring-wound heating wire.

[0012] Preferably, the front shell is sleeved on the external thread of the camera housing via an internal thread, and the rear shell is fixedly mounted on the camera housing via screws.

[0013] The present invention also provides a method for using a vehicle-mounted camera self-cleaning device based on ultrasonic vibration and scene classification, which is characterized by comprising the following steps: S1. Detect and identify weather scenes through the image acquisition detection and recognition module on the PCB circuit board. The module's weather scene recognition accuracy should reach more than 90%, and the misjudgment rate should be less than 10%. In rainy weather, increase the frequency of image acquisition and adjust the feature weights in the recognition algorithm to improve recognition accuracy. In foggy weather, use a special image enhancement algorithm to pre-process the collected images before identifying the weather scene.

[0014] S2. The piezoelectric transducer and the heating coil are connected respectively, and mechanical vibration is generated through the piezoelectric ceramic sheet inside the piezoelectric transducer and transmitted to the lens installed above, so that the lens shakes, thereby separating rain and fog droplets from the lens, and the lens is heated by connecting the heating coil, and then the heat is transferred to the lens, thereby accelerating the melting of frost into water droplets, and then the piezoelectric ceramic sheet generates mechanical vibration to shake off the water droplets, thereby completing the cleaning operation of the vehicle camera.

[0015] Preferably, the image acquisition detection and recognition module adopts an autonomous driving different weather scene classification algorithm based on deep learning, including input scene samples, feature extraction module, feature differentiation module, feature fusion understanding module and output different weather scene categories.

[0016] Preferably, the self-cleaning device includes the following hardware structure: A data storage device, used to store images captured by the vehicle camera or classification results obtained by the image receiving device; Program memory, used to call program instructions and execute deep learning-based autonomous driving classification algorithms for different weather scenarios; An interface device for providing an interface for wired or wireless communication; Communication device used for short or long range communication.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention integrates the ultrasonic vibration function into the vehicle-mounted camera, and realizes the ultrasonic vibration to remove rain, snow and mud while ensuring the image acquisition performance of the camera, so as to prevent the vehicle-mounted camera from being blocked by rain, snow, mud and sand during driving in different bad weather conditions, thereby affecting the driving safety of the car. At the same time, the present invention has the characteristics of simple structure, easy assembly, good reliability, miniaturization and low cost.

[0018] 2. The present invention also provides a method for using a self-cleaning device for a vehicle camera based on ultrasonic vibration and scene classification, which detects and identifies the current driving scene as rainy, foggy, snowy, sandstorm and other weather conditions, uses different cleaning strategies, heats the lens by connecting a heating coil, and then conducts heat to the lens, thereby accelerating the melting of frost into water droplets or evaporating moisture in mud and sand, and shakes off residual obstructions through ultrasonic vibration. The device can be applied to vehicle camera cleaning work in a variety of environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0020] In the attached picture: Figure 1 It is a schematic diagram of the overall structure of the present invention; Figure 2 is a schematic side cross-sectional view of the whole of the present invention; Figure 3 The present invention Figure 2 The enlarged schematic diagram of point A in the middle; Figure 4 The present invention Figure 2 The enlarged schematic diagram of point B in the middle; Figure 5 It is a schematic diagram of the classification algorithm of different weather scenarios for autonomous driving based on deep learning of the present invention; Figure 6 It is a schematic diagram of a feature extraction module of the present invention; Figure 7 It is a schematic diagram of the hardware structure of the ultrasonic vibration for removing rain, snow and mud of the vehicle-mounted camera of the present invention; Numbers in the figure: 1. lens; 2. camera housing; 3. camera assembly; 4. rear housing; 5. PCB circuit board; 6. front housing; 7. heating coil; 8. piezoelectric transducer; 9. piezoelectric ceramic sheet; 101. program memory; 102. data memory; 103. interface device; 104. communication device. DETAILED DESCRIPTION

[0021] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0022] Example: Figure 1-Figure 4 As shown, a self-cleaning device for a vehicle-mounted camera based on ultrasonic vibration and scene classification includes a lens 1, a camera assembly 3, a piezoelectric transducer 8, a camera housing 2, a PCB circuit board 5, a heating coil 7, a front housing 6 and a rear housing 4; the front housing 6 and the rear housing 4 are respectively located at the front and rear ends of the camera housing 2 and are fixedly connected, and a receiving cavity is formed therein after the camera housing 2 and the rear housing 4 are fixedly connected, and the camera assembly 3 is located in the receiving cavity, and a PCB circuit board 5 is fixedly installed in the camera assembly 3, and a control module is provided on the PCB circuit board 5, and a plug connector for electrically connecting to a vehicle body is fixedly connected to the rear housing 4, and the plug connector is electrically connected to the PCB circuit board 5, and can be used for powering the camera or transmitting information, etc.; the rear housing 4 is fixedly connected to the camera assembly 3 at the lower side of the PCB circuit board 5, and the rear housing 4 is fixedly installed with the camera housing 2 by screws, and the front housing 6 is screwed on the camera housing, and a mounting groove is provided in the camera housing 2, and a piezoelectric transducer 8 is fixedly connected to the mounting groove, and a piezoelectric ceramic is provided in the piezoelectric transducer 8 The piezoelectric ceramic sheet 9 is fixed with a lens 1, and the heating coil 7 is bonded around the lens 1. The heating coil 7 is an existing structure. The camera housing 2 is fixed with a connecting wire electrically connected to the piezoelectric transducer 8 and the heating coil 7, and the other end of the connecting wire is electrically connected to the PCB circuit board 5. When the lens 1 is covered with pollutants such as mud, sand, and frost, the heating coil 7 surrounds the lens 1 to heat the entire lens 1, and then slowly heats to melt the frost and evaporate the water without spraying water, jetting air, or directly heating the lens 1, thereby avoiding the problem of direct high temperature or water jet heating and abrasion of the lens 1 to damage the lens 1. The piezoelectric ceramic sheet 9 is bonded around the bottom of the lens 1. When water droplets, dust, melted frost, evaporated mud, and other pollutants are detected and identified on the lens 1, the pollutants are shaken out of the lens 1 through rapid ultrasonic vibration, and the vehicle-mounted camera can be used quickly and directly without spraying water, jetting air, or directly heating the lens 1, thereby avoiding the problem of direct high temperature or water jet heating and abrasion of the lens 1 to damage the lens 1.

[0023] Specifically, the rear shell 4 can be fixedly connected to the lower end of the camera housing 2 by means of screws, and the front shell 6 can be fixed to the front end of the camera housing 2 by means of threaded engagement.

[0024] Among them, Figure 4 As shown, the edge of the lens 1 is fixedly surrounded by a circle of heating coils 7, and the heating coils 7 can be set to a metal heat-conducting material. When the surface of the lens 1 is covered with pollutants such as frost, mud and sand, the lens 1 can be heated by the side heating coils 7, and then the heat is transferred to the lens 1, thereby melting the frost or mud on the surface of the lens 1, thereby better solving the problem that the surface of the lens 1 is covered with frost or mud and sand that cannot be completely removed by mechanical vibration alone.

[0025] Among them, Figure 5 and Figure 6 As shown, in a specific embodiment, scene samples are input to realize the process of classifying different weather scenes. The image taken by the vehicle camera is input, and the total feature map is output to the feature differentiation module through continuous convolution, CBAM, maximum pooling and other operations. Then, the feature map is output to the feature fusion classification module after feature extraction, feature differentiation and feature screening. In the feature fusion classification module, the input feature map is firstly reduced in dimension through average pooling and maximum pooling operations, and then the correlation between channels is learned through the RELU activation function of the fully connected layer respectively. Finally, the channel attention weight is generated through the fully connected layer and the sigmoid function, and finally the current weather scene category is output as one of rainy day, foggy day, snowy day or sandstorm weather.

[0026] like Figure 7 As shown, it is a hardware structure diagram of an ultrasonic vibration rain, snow and mud removal method of a vehicle-mounted camera provided in an embodiment of the present invention, including: an image acquisition device, an image receiving device and a cleaning implementation device, the image acquisition device is connected and communicated with the image receiving device by wired or wireless means to transmit relevant image data; the image acquisition device can receive images taken by the vehicle-mounted camera, and send them to the image receiving device of the receiving party through the interface device 103 and the communication device 104; the receiving party's device saves the received image in the data storage device 102, and determines the weather scene category of the current driving scene through the automatic driving different weather scene classification algorithm of this embodiment in the program storage device 101; the cleaning implementation device adopts a corresponding cleaning strategy according to the scene category output by the image receiving device, and the image receiving device may include Figure 7The computer program product may be a collection device and a receiving device, or other devices capable of image processing, which are not limited here; wherein, the computer program product contains a set of program instructions on a computer-readable storage medium, and these instructions are used to implement a specific invention or function. These instructions can exist in many forms, such as machine instructions, assembly language instructions, high-level language instructions, or scripting language instructions, and their selection depends on the design and application requirements of the program. A computer-readable medium is a physical device for storing these program instructions. It can be a hard disk, a random access memory (RAM), a read-only memory (ROM), or a portable storage device (such as a compact disk read-only memory). These media are used to store program instructions and provide the storage and access capabilities required by instruction execution devices (such as computers, microprocessors, or other execution devices).

[0027] The image acquisition device, the image receiving device and the cleaning implementation device all include a program memory 101, a data memory 102, an interface device 103 and a communication device 104, and are equipped with different types of memory and interface devices. The communication device supports multiple communication methods, including wired connection methods such as Ethernet, USB, HDMI and serial port, as well as wireless communication protocols such as Wi-Fi, Bluetooth and ZigBee. The program memory 101 of the image receiving device stores a computer program that can be used to perform specific operations; the program can be used to determine the current driving scene weather category and send it to the cleaning implementation device. The program memory 101 of the cleaning implementation device also stores a computer program for controlling the heating of the heating coil and the vibration of the piezoelectric ceramic piece to execute the corresponding cleaning strategy. The method can determine the current driving weather scene category by receiving the image taken by the vehicle camera, and execute the corresponding cleaning strategy to complete the function of cleaning the vehicle camera.

[0028] The present invention also provides a method for using a vehicle-mounted camera self-cleaning device based on ultrasonic vibration and scene classification, comprising the following steps: S1. Detect and identify the weather scene through the image acquisition detection and identification module on the PCB circuit board 5.

[0029] S2. The piezoelectric transducer 8 and the heating coil 7 are respectively turned on, and mechanical vibration is generated by the piezoelectric ceramic sheet 9 inside the piezoelectric transducer 8 and the mechanical vibration is transmitted to the lens 1 installed above, so that the lens 1 shakes, thereby separating the raindrops from the lens. At the same time, the lens 1 is heated by turning on the heating coil 7, and then the heat is transferred to the lens 1, thereby accelerating the melting of frost into water droplets, and then the piezoelectric ceramic sheet 9 generates mechanical vibration to shake off the water droplets.

[0030] This includes the following situations: 1) When the driving scene is rainy or foggy, the image receiving device on the PCB circuit board 5 detects and recognizes that the weather scene is rainy or foggy, and turns on the piezoelectric transducer 8. The piezoelectric ceramic piece 9 inside the piezoelectric transducer 8 generates mechanical vibration and transmits the mechanical vibration to the lens 1 installed above, causing the lens 1 to shake, thereby separating raindrops from the lens 1.

[0031] 2) When the driving scene is snowy, the image receiving device on the PCB circuit board 5 detects and recognizes that the weather scene is snowy, and turns on the heating coil 7 to heat the lens 1, thereby transferring the heat to the lens 1, thereby accelerating the melting of frost into water droplets, and then generating mechanical vibrations through the piezoelectric ceramic sheet 9 to shake off the water droplets.

[0032] 3) When the driving scene is a sandstorm, the image receiving device on the PCB circuit board 5 detects and recognizes that the weather scene is a sandstorm, and turns on the heating coil 7 to heat the lens 1, thereby evaporating the moisture in the mud and sand through heating, and shaking off the residual pollutants through mechanical vibration.

[0033] Finally, it should be noted that the above description is only a preferred example of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A self-cleaning device for a vehicle camera based on ultrasonic vibration and scene classification, characterized in that: The invention comprises a lens, a camera assembly, a piezoelectric transducer, a camera housing, a PCB circuit board, a heating coil, a front housing and a rear housing; the front housing and the rear housing are respectively fixedly connected at the front and rear ends of the camera housing, and a receiving cavity is formed therein after the camera housing is fixedly connected to the rear housing, the lens is fixed between the front housing and the camera housing, the piezoelectric transducer is fixed between the lower side of the lens and the groove provided in the camera housing, a piezoelectric ceramic sheet is provided inside the piezoelectric transducer, and the heating coil is installed on the inner side of the lens, the camera assembly is located inside the camera housing, a PCB circuit board is fixedly installed at the lower end of the camera assembly, the lower side of the PCB circuit board is fixedly connected to the rear housing, a plug connector for electrically connecting to a vehicle body is fixedly connected to the rear housing, the plug connector is electrically connected to the PCB circuit board, a connecting wire electrically connected to the piezoelectric transducer and the heating coil is fixedly connected to the camera housing, and the other end of the connecting wire is electrically connected to the PCB circuit board, the heating coil is wrapped around the lens to heat up the entire lens, and at the same time, the contaminants are shaken out of the lens range by wrapping around the piezoelectric ceramic sheet bonded below the lens.

2. The self-cleaning device for a vehicle camera based on ultrasonic vibration and scene classification according to claim 1, characterized in that: The piezoelectric transducer is directly fixed to the lower side of the lens by bonding, hot pressing or welding. The heating coil is installed on the inner side of the lens by bonding, hot pressing or welding and connected to the PCB circuit board. One end of the piezoelectric transducer uses a viscous material with thermal conductivity. After the piezoelectric transducer is wrapped around the camera housing, the two ends are bonded or hot pressed.

3. The self-cleaning device for a vehicle camera based on ultrasonic vibration and scene classification according to claim 1, characterized in that: Mechanical vibration is generated by the piezoelectric ceramic sheet, and the material of the piezoelectric ceramic sheet includes potassium sodium metaniobate or barium strontium metaniobate.

4. The self-cleaning device for a vehicle camera based on ultrasonic vibration and scene classification according to claim 3 is characterized in that: One side of the piezoelectric ceramic sheet is integrated with or uses a viscous material with heat conduction properties alone to achieve the fixation of the piezoelectric ceramic sheet and the lens.

5. The vehicle-mounted camera self-cleaning device based on ultrasonic vibration and scene classification according to claim 1, characterized in that: The heating coil obtains power and generates heat through the PCB circuit board, causing the lens to heat up. The heating coil is arranged around the inner ring of the lens barrel of the vehicle-mounted lens. The heating coil is in the form of a flexible circuit board or a ring-shaped winding heating wire.

6. The vehicle-mounted camera self-cleaning device based on ultrasonic vibration and scene classification according to claim 1, characterized in that: The front shell is sleeved on the external thread of the camera housing through an internal thread, and the rear shell is fixedly mounted on the camera housing through screws.

7. A method for using the self-cleaning device for a vehicle camera based on ultrasonic vibration and scene classification according to any one of claims 1 to 6, characterized in that: The following steps are involved: S1, detecting and identifying the weather scene through the image acquisition detection and recognition module on the PCB circuit board; S2. The piezoelectric transducer and the heating coil are connected respectively, and mechanical vibration is generated through the piezoelectric ceramic sheet inside the piezoelectric transducer and transmitted to the lens installed above, so that the lens shakes, thereby separating rain and fog droplets from the lens, and the lens is heated by connecting the heating coil, and then the heat is transferred to the lens, thereby accelerating the melting of frost into water droplets, and then the piezoelectric ceramic sheet generates mechanical vibration to shake off the water droplets, thereby completing the cleaning of the vehicle-mounted camera.

8. The method for using the vehicle camera self-cleaning device based on ultrasonic vibration and scene classification according to claim 7, characterized in that: The image acquisition detection and recognition module adopts an autonomous driving different weather scene classification algorithm based on deep learning, including input scene samples, feature extraction module, feature differentiation module, feature fusion understanding module and output different weather scene categories.

9. The method for using the vehicle camera self-cleaning device based on ultrasonic vibration and scene classification according to claim 8, characterized in that: The self-cleaning device includes the following hardware structures: A data storage device, used to store images captured by the vehicle camera or classification results obtained by the image receiving device; A program memory, used to call program instructions to execute the deep learning-based autonomous driving classification algorithm for different weather scenarios; An interface device for providing an interface for wired or wireless communication; Communication device used for short or long range communication.

Images