Low-power-consumption image crack recognition device and use method thereof

Abstract

The invention discloses a low-power-consumption image crack recognition device and a use method thereof. Comprising a low-power-consumption processor, an external RTC timing clock for triggering the low-power-consumption processor to be switched between a working mode and a standby mode, a peripheral module for realizing picture and image acquisition and transmission functions, and a power supply module for providing constant power for the low-power-consumption processor and the external RTC timing clock, the calibration system is used for realizing automatic calibration of an image picture; on the basis of an image recognition technology, a contact sensor does not need to be installed at a crack hidden danger position, the device only needs to be deployed at the position where a crack and a scale grid can be shot at the same time, data are collected through wireless transmission, secondary damage to a measured object due to the fact that expansion bolts, structural adhesives and other installation modes are used in traditional contact measurement is avoided, and the measurement accuracy is improved. Meanwhile, the integrity of the original hidden danger part is reserved, and the method is particularly suitable for monitoring infrastructures with structural hidden dangers.

Description

Technical Field

[0001] This invention belongs to the field of displacement monitoring technology, specifically relating to a low-power image crack recognition device and its usage method. Background Technology

[0002] Traditional crack monitoring of infrastructure such as bridges, tunnels, and slopes mostly uses contact methods for measurement. When using these methods, linear or wire-type crack gauges are installed at the crack locations, requiring rigid connections. Therefore, expansion bolts and structural adhesives are needed for installation, which can cause secondary damage to the original potential hazard points, and the measurement accuracy is not high enough.

[0003] Currently, there are also devices that use image and video principles to measure crack identification. The basic principle is to identify the size of the crack by the imaging features and pixel size of the crack in the video image. This method is a non-contact measurement, but the monitoring equipment requires a long-term power supply and usually uses wired communication, so the system is complicated to install, implement and maintain. In addition, some devices do not have online calibration functions, and the recognition error of the video image in the dark environment will affect the measurement results. Summary of the Invention

[0004] The purpose of this invention is to provide a low-power image crack recognition device and its usage method. Through low-power design, scale self-calibration, image filling calculation and physical highlighting process, it solves the problems of complex installation, high power consumption and measurement accuracy being easily affected in dark light environments of existing wired equipment.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A low-power image crack recognition device includes a low-power processor, an external RTC timer clock that triggers the low-power processor to switch between working mode and standby mode, a peripheral module for image acquisition and transmission, a power supply module that provides constant power to the low-power processor and the external RTC timer clock, and a calibration system for automatic image calibration; the low-power processor, the external RTC timer clock, the peripheral module power supply module, and the calibration system are all electrically connected. The calibration system includes a scale grid for quickly identifying the size and location of cracks, and a calculation module for pixel calculation. The calculation module acquires image information of the crack under test and the scale grid, and obtains the crack size by filling the crack image with the scale grid and comparing it.

[0006] Preferably, the scale grid has several alternating black and white grids, the grids are square, and the side length L is set to a range of 0.5mm-100mm.

[0007] Preferably, the calculation module includes an image preprocessing unit for converting the color image into a grayscale image and filtering out environmental noise in the image; The calibration unit is used to identify the size of the scale grid in the image, count the number of pixels N occupied by the known size scale grid in the image, and derive the actual size coefficient corresponding to a single pixel. The crack coarse measurement unit is used to locate the crack size range, make a preliminary judgment on whether it exceeds the set threshold, and narrow down the calculation range. The crack precision measurement unit, based on the coarse measurement results, calculates the actual size of the crack precisely. The threshold comparison and early warning triggering unit compares the calculated actual size of the crack with a preset threshold to trigger an early warning mechanism.

[0008] Preferably, the formula for calculating the actual size factor of a single pixel is L / N.

[0009] Preferably, the low-power processor is an HC32L196 model, the external RTC real-time clock is an RX8025SA model, and a backup button battery is provided for power backup.

[0010] Preferably, the peripheral module includes a serial port camera, an RS485 driver, and a 4GCAT1 communication device; The low-power processor is electrically connected to an external RTC real-time clock via an IIC interface, the UART port of the low-power processor is electrically connected to a serial camera, and the RS485 interface of the low-power processor is electrically connected to an RS485 driver.

[0011] Preferably, the serial port camera is 5 megapixels and supports TTL level, RS485 level or RS232 level output. The RS485 driver adopts the SP3485 driver to realize the system's compatibility support for TTL level and RS485 interface serial port cameras.

[0012] Preferably, the power supply module includes a lithium-ion battery, an LDO linear regulator with enable control, and a DC-DC boost module. The enable pin of the LDO linear regulator is connected to the PA07 pin of the low-power processor, and the DC-DC boost module boosts the lithium-ion battery from 3.7V to 12V to power the serial camera.

[0013] Preferably, the power supply module further includes a MOS switch for controlling the input of the DC-DC boost module, and the driving terminal of the MOS switch is connected to the PA06 pin of the low-power processor.

[0014] Based on the above description of a low-power image crack recognition device, the present invention also provides a method of using the low-power image crack recognition device, comprising the following steps: S1. Stick the ruler grid near the crack in the infrastructure being tested, apply fluorescent agent inside or outside the crack, or apply black and white paint around the crack, and install a low-power image crack recognition device with a serial port camera so that it can simultaneously capture the positions of the ruler grid and the crack being tested. S2. After the low-power processor is initialized upon first power-on, it configures the next working time to the external RTC real-time clock and then enters a low-power sleep state. S3. When the external RTC real-time clock reaches the set working time point, the low-power processor turns on the peripheral module power supply and sends an image capture command. The command specifies the pixel size of the captured image. The serial port camera captures an image containing the scale grid and the crack being measured according to the command and transmits the image data in packets to the computing module. S4. The actual size of the crack obtained by the calculation module is compared with the set threshold. If it exceeds the set threshold, the system triggers an early warning message and pushes it.

[0015] The low-power image crack recognition device and its usage method proposed in this invention have the following advantages compared with the prior art: 1. This invention, based on image recognition technology, eliminates the need to install contact sensors at the crack hazard site. Instead, the device is deployed at a location that can simultaneously capture images of the crack and the scale grid. Data is collected wirelessly, avoiding secondary damage to the object being measured caused by traditional contact measurement methods that use expansion bolts, structural adhesives, etc. At the same time, it preserves the integrity of the original hazard site, making it particularly suitable for monitoring infrastructure with existing structural hazards.

[0016] 2. This invention uses a black and white alternating square grid to capture images containing the grid and cracks, deriving the actual size coefficient of each pixel to achieve system self-calibration. Addressing the issue of low image recognition accuracy in low-light environments such as darkness, this invention employs a physical brightening process by applying fluorescent agents inside or outside the crack, or by applying black and white paint around it. This creates a significant color difference, enhancing image contrast. Combined with subsequent grayscale processing and noise reduction preprocessing steps, this ensures clear identification of the crack outline even in dark environments, resulting in more accurate measurement results.

[0017] 3. This invention maintains constant power supply through a low-power processor and RTC. Peripheral modules are only powered by the processor when working and are completely powered off when in hibernation, which can extend the battery life. It does not require external power supply or complex wiring and is suitable for long-term unattended scenarios such as bridges and tunnels. Attached Figure Description

[0018] Figure 1 This is a system block diagram of the present invention; Figure 2 This is a schematic diagram showing the relative positions of the crack and the scale grid when the device of the present invention is applied. Figure 3 This is a schematic diagram of the scale grid of the present invention; Figure 4 This is a circuit diagram of the MCU power supply circuit in the power supply module of the present invention; Figure 5 This is a power supply circuit diagram for the peripheral module in the power supply module of the present invention; Figure 6 This is a connection diagram of the enable pin of the MT3608 chip in this invention; Figure 7 This is a diagram of the DC-DC boost circuit of the present invention; Figure 8 This is a circuit diagram of the external RTC timing clock of the present invention; Figure 9 This is the circuit diagram of the serial port camera of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] This invention provides, for example Figure 1-9 The low-power image crack recognition device shown includes a low-power processor, an external RTC timer clock that triggers the low-power processor to switch between working mode and standby mode, a peripheral module for image acquisition and transmission, a power supply module that provides constant power to the low-power processor and the external RTC timer clock, and a calibration system for automatic image calibration. The low-power processor, the external RTC timer clock, the peripheral module power supply module, and the calibration system are all electrically connected. The power supply enable control of the peripheral module is controlled by the start / stop control of the low-power processor. When the device is in sleep mode, the peripheral is powered off, and when it is in operation, the power supply is turned on, thereby reducing the power consumption of the system in standby mode and improving the battery life. The calibration system includes a scale grid for quickly identifying the size and location of cracks, and a calculation module for pixel calculation. The calculation module acquires image information of the crack under test and the scale grid, and obtains the crack size by filling the crack image with the scale grid and comparing it.

[0021] scale grid Figure 3As shown, the scale grid consists of several alternating black and white grids. Each grid is square, with a side length L ranging from 0.5mm to 100mm. The scale grids are 0.5mm*0.5mm, 1mm*1mm, 2mm*2mm, 3mm*3mm, 4mm*4mm, and so on. These different sizes are designed for rapid, rough identification of crack size, followed by finer judgment based on pixel calculations. The scale grid is attached and fixed near the object being measured. During image capture, the serial camera simultaneously captures both the scale grid and the crack. Calculations are performed by obtaining the pixel value N occupied by the known-sized scale grid. The size of a single pixel is then L / N. After obtaining the crack's pixel information, this coefficient is used to calculate the actual crack size. The square grid is used because both vertical and horizontal pixels can be used for calibration reference.

[0022] The calculation module includes an image preprocessing unit, which converts color images into grayscale images and filters out environmental noise in the images; by converting color images into grayscale images, the amount of data calculation is simplified.

[0023] The calibration unit is used to identify the size of the scale grid in the image, extract its known side length L, count the number of pixels N occupied by the known size scale grid in the image, and derive the actual size coefficient corresponding to a single pixel as the benchmark for subsequent crack size conversion. The crack coarse measurement unit is used to locate the crack size range, initially determine whether it exceeds a set threshold, and narrow down the calculation range. Coarse measurement is mainly achieved through the following methods: The preprocessed image is converted into black and white binary data to determine the boundaries between the crack, the background, and the scale grid. The comparison sample includes scale grid data and the measured crack data. The scale grid acts like a scale, which is advantageous because it has different sizes of known dimensions. Using this control group, the pixel width data of the crack is bubble sorted with the pixel data of the scale grid of different sizes. The sorting results determine the two adjacent scale grid intervals where the crack size is located. For example, if the crack size is between the 2mm×2mm and 3mm×3mm scale grids, coarse localization is achieved. In this way, the size of the crack can be quickly and coarsely located.

[0024] After coarse measurement, fine measurement is performed. The crack fine measurement unit calculates the actual size of the crack based on the coarse measurement results. The formula for calculating the actual size coefficient of a single pixel is L / N. Precise calculation mainly includes two calculation methods: Fill-accumulation algorithm: The scale grid and the image of the crack under test are finely processed. The crack image is filled and compared with the scale grid. The crack image is filled 1:1 with scale grids smaller than the crack size. The number of complete scale grids filled is counted. Then the pixel value of the remaining part is calculated. Since the size of the scale grid is known, the pixel value N occupied by the known size scale grid is obtained. Then the size of a single pixel is L / N. The number of extra pixel grids * L / N is the size of the extra part. Adding the size of the already filled scale grid is the size of the crack. Direct pixel conversion algorithm: Directly count the total number of pixels M occupied by the crack image. Since the size of the scale grid is known, the crack size is directly calculated by obtaining the pixel value N occupied by the known size scale grid. Actual size = M × (L / N). The threshold comparison and early warning triggering unit compares the calculated actual crack size with a preset threshold. If the crack size exceeds the threshold, an early warning message is generated and pushed to the monitoring platform via the 4GCAT1 communication device to trigger the early warning mechanism.

[0025] like Figure 8 As shown, the low-power processor is an HC32L196 model, and the external RTC real-time clock is an RX8025SA model. The power supply is shared with the low-power processor and is also a constant power supply. It is also equipped with a spare button battery for power backup to prevent data loss during power failure.

[0026] The system is powered by a 3.7V lithium-ion battery. The advantages of lithium-ion batteries are high energy density and low self-discharge rate, making them suitable for low-power products and low-power processors.

[0027] like Figure 5 and Figure 6 As shown, the peripheral module includes a serial port camera, an RS485 driver, and a 4GCAT1 communication device; the power supply for other peripherals is mainly divided into two paths: DVCC_+3.3V and VCC_+12V, where DVCC_+3.3V uses an LDO linear regulator. Figure 7 As shown, VCC_+12V is implemented using the MT3608 chip.

[0028] like Figure 4 and Figure 5As shown, the power supply module includes a lithium-ion battery, an LDO linear regulator with enable control, and a DC-DC boost module. The enable pin of the LDO linear regulator is connected to the PA07 pin of the low-power processor to achieve switching control. The LDO linear regulator uses an ME6339C33M5G to reduce the battery voltage to 3.3V, and the enable pin of the LDO linear regulator is connected to the positive terminal of the input voltage, so there is always a 3.3V output. Since the low-power processor is powered by constant power, it is always powered. The DC-DC boost module boosts the lithium-ion battery voltage from 3.7V to 12V to power the serial port camera.

[0029] The power supply module also includes a MOS switch for controlling the input of the DC-DC boost module, and the driving terminal of the MOS switch is connected to the PA06 pin of the low-power processor. When the power supply of DVCC_+3.3V and VCC_+12V needs to be turned on, the low-power processor sets the PA06 and PA07 pins to a high level to turn them on and sets them to a low level to turn them off, thereby realizing the switching control of the power supply. The low-power processor is electrically connected to an external RTC real-time clock via an IIC interface, the UART port of the low-power processor is electrically connected to a serial camera, and the RS485 interface of the low-power processor is electrically connected to an RS485 driver.

[0030] The device operates in two states: active and sleep. After initial power-on initialization, the low-power processor obtains the next operating time and configures the alarm settings to the RTC chip via the IIC interface. After configuration, the processor enters a low-power sleep state with a standby current of less than 10μA, and the device is in sleep mode most of the time. While the processor is in sleep mode, the external RTC real-time clock continues to keep time. When the RTC clock reaches the required operating time, its interrupt pin generates an interrupt signal, waking the processor. The external RTC real-time clock communicates and configures with the low-power processor via the low-power processor's IIC1 interface, and the alarm interrupt pin of the external RTC real-time clock is connected to the low-power processor's interrupt PB03 to implement the alarm interrupt function.

[0031] The serial port camera is 5 megapixels and supports TTL, RS485, or RS232 level output. The RS485 driver uses an SP3485 driver, ensuring system compatibility with TTL and RS485 interface serial port cameras. This serial port camera replaces traditional cameras, avoiding high integration complexity and high power consumption, making it suitable for long-term structural monitoring scenarios.

[0032] The image sensor implementation scheme uses traditional cameras with output interfaces such as Ethernet and USB, primarily for video image transmission. Considering the low transmission frequency (once a day for data acquisition and transmission) and low power consumption requirements for structural crack monitoring, this application uses a 5-megapixel serial camera as the image capturing device. This allows for integration using a low-power microcontroller. The serial camera can be TTL, RS485, or RS232 level.

[0033] The image sensor operates as follows: a low-power processor sends image capture commands via a UART port, specifying the image size and pixel resolution, such as 2560*1920 or 1920*1080. After the serial camera captures the image, it transmits the image as a hexadecimal file, packetized and sent to the processor via the UART port. The processor then transmits the data in packets via a 4GCAT1 module, with each packet distinguished by its header sequence number. Once the entire image is transmitted, one acquisition and reading operation is complete. After the image data is transmitted to the platform in hexadecimal packets, the platform integrates the data according to a report and displays the image in JPEG format, thus enabling the microcontroller to drive the camera.

[0034] like Figure 9 As shown, this application uses the SP3485 driver to implement RS485 interface communication. This design can achieve system compatibility and supports serial cameras with TTL level and RS485 interface.

[0035] Based on the above description of a low-power image crack recognition device, the present invention also provides a method of using the low-power image crack recognition device, comprising the following steps: S1. Attach and fix the scale grid near the cracks in the infrastructure being measured, such as... Figure 2 As shown, the shaded area represents the crack. Applying fluorescent agent to the inside or outside of the crack or applying black and white paint around the crack enhances the visibility of the crack edge and strengthens the difference in brightness between the crack and the background, and between the scale grid and the image. Installing a low-power image crack recognition device with a serial port camera can simultaneously capture the positions of the scale grid and the crack being measured. S2. Upon initial power-on, the low-power processor automatically initiates the initialization process: completing hardware interface self-tests, including connection verification with the external RTC real-time clock, serial camera, RS485 driver, and 4GCAT1 communication device; loading system parameters including preset shooting pixel specifications, data transmission baud rate, and threshold standards. After initialization, the processor communicates with the external RTC real-time clock via the IIC interface, configuring the next working time into the RTC chip and setting the RTC alarm interrupt trigger mode. After configuration, the processor automatically disables power supply to unnecessary hardware and enters a deep sleep low-power state. At this time, the system standby current can be controlled below 10μA, reducing battery power consumption and extending device battery life. S3. When the external RTC real-time clock reaches the set working time, its alarm interrupt pin will generate a high-level interrupt signal, waking up the low-power processor from sleep mode via hardware connection. After the processor wakes up, it immediately outputs a high-level signal through PA06 and PA07 pins to control the MOS switch and LDO linear regulator to turn on, providing stable power to the peripheral modules: the DC-DC boost module boosts the 3.7V voltage of the lithium-ion battery to 12V to power the serial camera, and the LDO linear regulator regulates the voltage to 3.3V to power the RS485 driver and 4GCAT1 module; After the power supply is stable, the processor sends an image capture command to the serial camera via the UART port. The command specifies the pixel size of the captured image. The serial camera captures an image containing the scale grid and the crack being measured, and then transmits the image data in packets to the computing module. S4. After receiving the complete image data packet, the calculation module starts a multi-step data processing flow: First, the color image is converted into a grayscale image through the image preprocessing unit, the Gaussian filtering algorithm is used to filter out environmental interference noise, and the contrast enhancement algorithm is used to further enhance the boundary differences between the crack and the background and the scale grid. Subsequently, the calibration unit automatically identifies the known side length L of the scale grid in the image, counts the number of pixels N it occupies in the image, and derives the size conversion benchmark by using the formula: actual size coefficient of a single pixel = L / N. Then, the crack coarse measurement unit converts the preprocessed image into black and white binary data, and compares the pixel width data of the crack with the pixel data of different size scale grids using the bubble sort algorithm to quickly locate the adjacent scale grid interval where the crack size is located, preliminarily determine whether it exceeds the set threshold, and narrow down the subsequent fine calculation range. Finally, based on the coarse measurement results, the crack fine measurement unit calculates the actual crack size using either a fill-accumulation algorithm or a direct pixel conversion algorithm: the fill-accumulation algorithm fills the crack image 1:1 with scale grids smaller than the crack size, counts the number of complete filled grids and calculates the pixel size of the remaining part, and accumulates them to obtain the total size; the direct pixel conversion algorithm directly counts the total number of pixels M occupied by the crack and calculates the accurate result using the formula actual size = M × (L / N). After the calculation is completed, the threshold comparison and early warning triggering unit compares the actual size of the crack with the preset threshold in real time. If the actual size of the crack does not exceed the threshold, the system records and stores the measurement data, and then the processor controls the peripheral module to power off and the system re-enters sleep mode. If the actual size of the crack exceeds the preset threshold, the system immediately generates an early warning report containing information such as crack location, actual size, measurement time, and device number, and pushes the early warning information to the remote monitoring platform through the 4GCAT1 communication device.

[0036] Finally, it should be noted that the above description is only a preferred embodiment 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 foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A low-power image crack recognition device, characterized in that: The system includes a low-power processor, an external RTC timer clock that triggers the low-power processor to switch between working mode and standby mode, a peripheral module for image acquisition and transmission, a power supply module that provides constant power to the low-power processor and the external RTC timer clock, and a calibration system for automatic image calibration; the low-power processor, the external RTC timer clock, the peripheral module power supply module, and the calibration system are all electrically connected. The calibration system includes a scale grid for quickly identifying the size and location of cracks, and a calculation module for pixel calculation. The calculation module acquires image information of the crack under test and the scale grid, and obtains the crack size by filling the crack image with the scale grid and comparing it.

2. The low-power image crack recognition device according to claim 1, characterized in that: The scale grid has several alternating black and white grids, each grid being square, and the side length L is set to range from 0.5mm to 100mm.

3. The low-power image crack recognition device according to claim 2, characterized in that: The calculation module includes an image preprocessing unit, used to convert color images into grayscale images and filter environmental noise in the images; The calibration unit is used to identify the size of the scale grid in the image, count the number of pixels N occupied by the known size scale grid in the image, and derive the actual size coefficient corresponding to a single pixel. The crack coarse measurement unit is used to locate the crack size range, make a preliminary judgment on whether it exceeds the set threshold, and narrow down the calculation range. The crack precision measurement unit, based on the coarse measurement results, calculates the actual size of the crack precisely. The threshold comparison and early warning triggering unit compares the calculated actual size of the crack with a preset threshold to trigger an early warning mechanism.

4. The low-power image crack recognition device according to claim 3, characterized in that: The formula for calculating the actual size factor of a single pixel is L / N.

5. The low-power image crack recognition device according to claim 1, characterized in that: The low-power processor is an HC32L196 model, and the external RTC real-time clock is an RX8025SA model, and is equipped with a backup button battery for power supply backup.

6. The low-power image crack recognition device according to claim 5, characterized in that: The peripheral module includes a serial port camera, an RS485 driver, and a 4GCAT1 communication device. The low-power processor is electrically connected to an external RTC real-time clock via an IIC interface, the UART port of the low-power processor is electrically connected to a serial camera, and the RS485 interface of the low-power processor is electrically connected to an RS485 driver.

7. The low-power image crack recognition device according to claim 6, characterized in that: The serial port camera is 5 megapixels and supports TTL, RS485, or RS232 level output. The RS485 driver uses the SP3485 driver to achieve system compatibility support for TTL level and RS485 interface serial port cameras.

8. The low-power image crack recognition device according to claim 7, characterized in that: The power supply module includes a lithium-ion battery, an LDO linear regulator with enable control, and a DC-DC boost module. The enable pin of the LDO linear regulator is connected to the PA07 pin of the low-power processor. The DC-DC boost module boosts the lithium-ion battery from 3.7V to 12V to power the serial port camera.

9. A low-power image crack recognition device according to claim 8, characterized in that: The power supply module also includes a MOS switch for controlling the input of the DC-DC boost module, and the driving terminal of the MOS switch is connected to the PA06 pin of the low-power processor.

10. A method of using a low-power image crack recognition device according to any one of claims 1-9, characterized in that: Includes the following steps: S1. Stick the ruler grid near the crack in the infrastructure being tested, apply fluorescent agent inside or outside the crack, or apply black and white paint around the crack, and install a low-power image crack recognition device with a serial port camera so that it can simultaneously capture the positions of the ruler grid and the crack being tested. S2. After the low-power processor is initialized upon first power-on, it configures the next working time to the external RTC real-time clock and then enters a low-power sleep state. S3. When the external RTC real-time clock reaches the set working time point, the low-power processor turns on the peripheral module power supply and sends an image capture command. The command specifies the pixel size of the captured image. The serial port camera captures an image containing the scale grid and the crack being measured according to the command and transmits the image data in packets to the computing module. S4. The actual size of the crack obtained by the calculation module is compared with the set threshold. If it exceeds the set threshold, the system triggers an early warning message and pushes it.

Images