Field monitoring system and method based on image accumulation

Abstract

This application provides a field monitoring system and method based on image accumulation, relating to the field of image processing technology. The system includes: a microcontroller for acquiring the analog voltage value output by a thermal infrared sensor, and switching the power-off sleep and power-on working states of the acquisition unit according to the analog voltage value and a voltage trigger threshold; the acquisition unit for capturing and storing images in the power-on working state, and sending a full-load cleaning request to the microcontroller when the number of captured images reaches a batch processing threshold; the microcontroller for receiving the full-load cleaning request and controlling the computing unit to switch from power-off sleep to power-on working state; and the computing unit for batch processing the stored captured images in the power-on working state, generating a sensitivity adjustment command based on the no-shot rate and sending it to the microcontroller, so that the microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the ambient temperature value. This can effectively reduce start-stop energy consumption and significantly extend field endurance.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a field monitoring system and method based on image accumulation. Background Technology

[0002] With the development of the Internet of Things and ecological monitoring technology, infrared monitoring equipment in the wild, such as infrared cameras, has become a powerful tool for wildlife surveys and ecological protection. Its core requirement is to achieve accurate monitoring and data collection of target animals with low power consumption in long-term unattended wild environments.

[0003] In related technologies, most outdoor infrared monitoring systems employ fixed-sensitivity thermal infrared sensors. Their triggering depends on the temperature difference between the target and the environment, making them susceptible to interference from complex outdoor environments. Furthermore, to achieve intelligent image selection, existing systems often use a single-trigger, single-wake-up, single-image-processing approach. Since the energy consumption of the computing unit startup process is far higher than the energy consumption of processing a single image, frequent startups result in a significant waste of battery energy in meaningless startup waiting times, failing to meet the low-power consumption requirements of long-term outdoor deployments. Therefore, ensuring the reliability of target acquisition while maximizing system endurance is crucial. Summary of the Invention

[0004] This application provides a field monitoring system and method based on image accumulation.

[0005] According to a first aspect of this application, a field monitoring system based on image accumulation is provided, the system comprising:

[0006] Microcontroller, acquisition unit, computing unit and sensing unit;

[0007] The sensing unit includes a thermal infrared sensor and a temperature sensor. The thermal infrared sensor is used to convert the sensed changes in infrared radiation into an analog voltage value, and the temperature sensor is used to collect the current ambient temperature value.

[0008] The microcontroller is used to acquire the analog voltage value output by the thermal infrared sensor, and to switch the power-off sleep state and the power-on working state of the acquisition unit according to the relationship between the analog voltage value and the voltage trigger threshold.

[0009] The acquisition unit is used to capture and store images locally when it is powered on, and to send a full-load cleaning request to the microcontroller when the number of captured images reaches the batch processing threshold.

[0010] The microcontroller is also used to control the computing unit to switch from a power-off sleep state to a power-on working state when it receives the full-load cleaning request.

[0011] The computing unit is used to perform batch processing on stored captured images when the power is on. It generates a sensitivity adjustment command based on the no-shot rate of the captured images in this batch processing and sends it to the microcontroller, so that the microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the current ambient temperature value.

[0012] Optionally, the acquisition unit is used to send an acquisition end command to the microcontroller if the number of image captures has not reached the batch processing threshold;

[0013] The microcontroller is used to control the acquisition unit to switch from a powered-on working state to a powered-off sleep state when it receives the acquisition end command.

[0014] Optionally, the computing unit is used to perform target detection processing on the stored captured images to identify target images containing animal data and invalid images that do not contain animal data, determine the no-shot rate based on the number of invalid images and a batch processing threshold, and delete the invalid images.

[0015] Optionally, the sensitivity adjustment command includes a sensitivity downgrade command and a sensitivity upgrade command;

[0016] The computing unit is used to generate a sensitivity downgrade instruction when the empty shot rate is greater than a first empty shot threshold, and to generate a sensitivity upgrade instruction when the empty shot rate is less than a second empty shot threshold.

[0017] Optionally, the computing unit is used to detect the number of currently unprocessed captured images after generating the sensitivity adjustment instruction, and send a power-off request to the microcontroller if the number of unprocessed captured images is less than the batch processing threshold.

[0018] Optionally, upon receiving the sensitivity adjustment instruction, the microcontroller searches the mapping data table based on the blanking rate to determine the corresponding target cumulative variable, and processes the reference voltage trigger threshold, the ambient temperature value, the preset compensation coefficient, and the target cumulative variable to update the voltage trigger threshold.

[0019] Optionally, the microcontroller is configured to determine the minimum boundary threshold as the voltage trigger threshold when the voltage trigger threshold is less than the minimum boundary threshold, and to determine the maximum boundary threshold as the voltage trigger threshold when the voltage trigger threshold is greater than the maximum boundary threshold.

[0020] Optionally, the microcontroller is used to perform a power-off cold start process on the computing unit when it controls the computing unit to enter the power-on working state and does not receive a response signal from the computing unit within a preset time.

[0021] Optionally, the field monitoring system includes a power management unit;

[0022] The power management unit includes a first switching device and a second switching device;

[0023] The first switching device is used to turn on when the microcontroller outputs a high-level signal at its first terminal, so that the acquisition unit enters a powered-on working state, and to turn off when the microcontroller outputs a low-level signal, so that the acquisition unit enters a powered-off sleep state.

[0024] The second switching device is used to turn on when the microcontroller outputs a high-level signal at its second terminal, so that the computing unit enters a powered-on working state, and to turn off when the microcontroller outputs a low-level signal, so that the computing unit enters a power-off sleep state.

[0025] According to a second aspect of this application, a field monitoring method based on image accumulation is provided, applicable to any of the image accumulation-based field monitoring systems described above, comprising:

[0026] The microcontroller switches between the power-off sleep state and the power-on working state of the acquisition unit based on the relationship between the analog voltage value and the voltage trigger threshold. The analog voltage value is the analog voltage value output by the thermal infrared sensor after converting the infrared radiation changes it senses.

[0027] When the acquisition unit is powered on, it captures and stores images locally, and sends a full-load cleaning request to the microcontroller when the number of captured images reaches the batch processing threshold.

[0028] Upon receiving the full-load cleaning request, the microcontroller controls the computing unit to switch from a power-off sleep state to a power-on working state.

[0029] When the computing unit is powered on, it performs batch processing on the stored captured images, generates a sensitivity adjustment command based on the no-shot rate of the captured images in this batch processing, and sends it to the microcontroller.

[0030] The microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the current ambient temperature value.

[0031] In summary, the field monitoring system based on image accumulation provided in this application has at least the following beneficial effects: The field monitoring system based on image accumulation includes a microcontroller, an acquisition unit, a computing unit, and a sensing unit. The microcontroller switches the acquisition unit between a power-off sleep state and a power-on working state based on the relationship between the analog voltage value output by the thermal infrared sensor and the voltage trigger threshold. When the acquisition unit is in the power-on working state, it performs image capture and local storage. When the number of captured images reaches the batch processing threshold, it sends a full-load cleaning request to the microcontroller. Upon receiving the full-load cleaning request, the microcontroller controls the computing unit to switch from the power-off sleep state to the power-on working state. When the computing unit is in the power-on working state, it performs batch processing on the stored captured images, generates a sensitivity adjustment command based on the no-shot rate, and sends it to the microcontroller so that the microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the current ambient temperature value. Therefore, by working together with the sensing unit, microcontroller, acquisition unit and computing unit, and combining image accumulation and batch processing mechanisms, the energy consumption of the computing unit can be effectively reduced during start-up and shutdown, and the endurance in the field can be greatly extended. At the same time, by dynamically adjusting the voltage trigger threshold of the acquisition unit in combination with the ambient temperature and the no-shot rate, false triggers caused by environmental interference can be effectively reduced, and the missed monitoring of target animals can be avoided. This significantly improves the reliability of ecological monitoring in unattended field scenarios and fully meets the diverse field monitoring needs. Attached Figure Description

[0032] To more clearly illustrate the specific embodiments of this application or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 A schematic diagram of a field monitoring system based on image accumulation provided for an embodiment of this application;

[0034] Figure 2 A schematic diagram of another field monitoring system based on image accumulation provided for an embodiment of this application;

[0035] Figure 3 A flowchart illustrating a field monitoring method based on image accumulation, provided for embodiments of this application;

[0036] Figure 4 A schematic diagram of a field monitoring process based on image accumulation, provided for an embodiment of this application;

[0037] Figure 5This is a schematic diagram of the internal logic of a primary control unit provided in an embodiment of this application. Detailed Implementation

[0038] To make the above and other features and advantages of this application clearer, the application is further described below with reference to the accompanying drawings. It should be understood that the specific embodiments given herein are for the purpose of explanation to those skilled in the art, and are exemplary only, not restrictive.

[0039] In the following description, numerous specific details are set forth to provide a thorough understanding of this application. However, it will be apparent to those skilled in the art that the specific details are not required to practice this application. In other instances, well-known steps or operations have not been described in detail to avoid obscuring this application.

[0040] refer to Figure 1 This application provides a field monitoring system based on image accumulation, which includes a microcontroller, an acquisition unit, a computing unit, and a sensing unit.

[0041] The sensing unit includes a thermal infrared sensor and a temperature sensor. The thermal infrared sensor is used to convert the sensed changes in infrared radiation into analog voltage values, and the temperature sensor is used to collect the current ambient temperature value.

[0042] Among them, the thermal infrared sensor (PIR), also known as a passive infrared detection sensor, typically contains a pyroelectric element. When a moving target, such as an animal, enters the detection range, the radiated infrared energy can be sensed by the pyroelectric element, causing a change in the element's charge distribution. After processing by the internal amplification circuit, a continuous analog voltage signal can be output. Therefore, in the embodiments of this application, the thermal infrared sensor can be used to monitor target animals within the range.

[0043] In addition, the microcontroller is used to acquire the analog voltage value output by the thermal infrared sensor, and to switch the power-off sleep state and the power-on working state of the acquisition unit according to the relationship between the analog voltage value and the voltage trigger threshold.

[0044] The microcontroller can be understood as the primary control unit in this field monitoring system. It typically has low power consumption and can integrate an analog-to-digital converter (ADC) interface, a general-purpose input / output (GPIO) interface, and a serial communication module. It can be used for environmental data acquisition, equipment power management, and command issuance.

[0045] When the field monitoring system is initially started, the voltage trigger threshold can be a pre-set reference voltage trigger threshold, or it can be pre-compensated based on the current ambient temperature. For example, if the current ambient temperature is greater than a first temperature threshold, the reference voltage trigger threshold can be lowered; conversely, if the ambient temperature is less than a second temperature threshold, the reference voltage trigger threshold can be increased.

[0046] For example, if the first temperature threshold is 30°C, and the current ambient temperature exceeds 30°C, the reference voltage trigger threshold can be lowered; or if the second temperature threshold is 10°C, and the current ambient temperature is less than 10°C, the reference voltage trigger threshold can be increased, etc. This application does not limit this.

[0047] Therefore, in this embodiment, the reference voltage trigger threshold can be pre-compensated based on the current ambient temperature value. When the ambient temperature is too high, the reference voltage trigger threshold can be reduced to improve the sensitivity of the thermal infrared sensor and avoid missed detections caused by the small temperature difference between the target and the environment. When the ambient temperature is too low, the reference voltage trigger threshold can be increased to reduce the sensitivity of the thermal infrared sensor and effectively suppress false triggering caused by interference such as thermal noise in low-temperature environments.

[0048] Understandably, the microcontroller can compare the analog voltage value output by the thermal infrared sensor in real time with the voltage trigger threshold after temperature pre-compensation. If the analog voltage value output by the thermal infrared sensor is greater than the voltage trigger threshold, it indicates that there is a change in infrared radiation of a suspected target animal within the monitoring range. The microcontroller can control the acquisition unit to switch from a power-off sleep state to a power-on working state. If the analog voltage value output by the thermal infrared sensor is less than or equal to the voltage trigger threshold, it indicates that there is no effective target at present. The microcontroller then maintains the power-off sleep state of the acquisition unit, or controls the acquisition unit to return to the power-off sleep state after completing a single acquisition.

[0049] Therefore, in this embodiment, the microcontroller can realize on-demand power-on and invalid sleep of the acquisition unit based on the relationship between the analog voltage value output by the thermal infrared sensor and the voltage trigger threshold. This reduces redundant image acquisition caused by invalid triggers from the source, ensures the reliability of target acquisition under different ambient temperatures through temperature pre-compensation, and minimizes energy consumption in non-working state, providing conditions for further optimization of battery life by waking up the computing unit later.

[0050] In addition, the acquisition unit can be used to capture and store images locally while in a powered-on state, and send a full-load cleaning request to the microcontroller when the number of captured images reaches the batch processing threshold.

[0051] The batch processing threshold can be a pre-set value, such as 20, 35, 40, etc., or it can be adjusted according to the actual needs such as the model and specifications of the acquisition unit, the frequency of animal activity in the wild environment, and the processing performance of the computing unit. This application does not limit this.

[0052] In addition, the full-load cleaning request can be used to indicate that the number of captured images currently acquired and stored by the acquisition unit has reached the preset batch processing threshold. This can trigger the microcontroller to wake up the computing unit, switch the computing unit from the power-off sleep state to the power-on working state, and perform image batch processing operations.

[0053] In addition, the acquisition unit can serve as a secondary acquisition unit in a field monitoring system. When powered on, it initiates image acquisition, capturing images according to preset shooting parameters such as resolution and frame rate, and storing the captured images in real time to a preset local storage area, such as shared storage, built-in flash memory, or external storage modules. Simultaneously, the internal counting unit of the acquisition unit can count the number of captured and stored images in real time. When the counted number of captured images reaches a preset batch processing threshold, a full-load cleaning request is sent to the microcontroller via the communication interface to trigger the batch image processing flow of subsequent computing units.

[0054] Optionally, the acquisition unit can also be used to send an acquisition end command to the microcontroller when the number of image captures has not reached the batch processing threshold. The microcontroller can then control the acquisition unit to switch from a power-on working state to a power-off sleep state upon receiving the acquisition end command.

[0055] For example, if the batch processing threshold is 30, and the number of captured images is 10, which is less than the threshold, the acquisition unit can send a completion command to the microcontroller. The microcontroller can then cut off the power to the acquisition unit, switching it from a powered-on state to a power-off sleep state. Alternatively, if the number of captured images is 31, which is greater than the threshold, the acquisition unit can send a full-load cleaning request to the microcontroller, which can then wake up the computing unit for subsequent processing.

[0056] It should be noted that the above examples are merely illustrative and should not be taken as limitations on the batch processing threshold, the number of captured images, etc. in the embodiments of this application.

[0057] Optionally, when the acquisition unit is powered on, it accumulates the count of captured and stored images. When the count reaches the batch processing threshold, it can send a full-load cleaning request to the microcontroller. If there is still a need for new image capture, such as if the thermal infrared sensor still detects a valid target, the acquisition unit can reset the original accumulated count to zero while sending the full-load request, and use the newly captured image as the first image in the new counting cycle for counting. Thus, in this embodiment, the count reset rule can ensure that each round of batch processing corresponds to image data within a fixed threshold range, avoiding the number of images processed by the computing unit at one time due to the accumulation of counts. At the same time, it ensures that the acquisition unit can continue to acquire the next round of images and count independently while the computing unit is processing the current round of images, without interrupting target acquisition, and can trigger the batch processing process of the computing unit.

[0058] In addition, the microcontroller can also be used to control the computing unit to switch from a power-off sleep state to a power-on working state when a full-load cleaning request is received. In the power-on working state, the computing unit can perform batch processing of stored captured images, generate sensitivity adjustment instructions based on the no-shot rate of the captured images in this batch processing, and send them to the microcontroller. This allows the microcontroller to adjust the voltage trigger threshold according to the sensitivity adjustment instructions and the current ambient temperature.

[0059] The computing power unit can be understood as the third-level computing power unit in the field monitoring system. It can be a Raspberry Pi, or other computing power chips or modules, etc. This application does not limit it.

[0060] In addition, the computing unit can load the target detection model after startup and read N accumulated captured images from the storage area at one time for inference, where N is the batch processing threshold.

[0061] Furthermore, the target detection model can be any network model capable of target detection, such as the YOLO model, the MobileNetV3 small model, or any other image classification model, etc. This application does not limit it in this regard.

[0062] Optionally, the computing unit can be used to perform target detection processing on the stored captured images to identify target images containing animal data and invalid images that do not contain animal data, determine the no-shot rate based on the number of invalid images and the batch processing threshold, and delete invalid images.

[0063] The computing unit can call the target detection model and input N captured images into the target monitoring model. After processing by the target monitoring model, it can output the category label and confidence level corresponding to each captured image. Then, based on the preset confidence level threshold, it can further determine the target images containing animal data and the invalid images that do not contain animal data.

[0064] After image classification is completed, the ratio of the number of invalid images to the batch processing threshold can be determined as the no-shot rate. Invalid images are automatically deleted, while target images containing animals are retained, thereby reducing the storage space occupied by redundant data.

[0065] Optionally, the sensitivity adjustment command may include a sensitivity downgrade command and a sensitivity upgrade command. The computing unit can generate a sensitivity downgrade command when the blank shot rate is greater than a first blank shot threshold, and generate a sensitivity upgrade command when the blank shot rate is less than a second blank shot threshold.

[0066] The first empty-shot threshold is greater than the second empty-shot threshold. The first and second empty-shot thresholds can be pre-set values, such as the first empty-shot threshold being 80% and the second empty-shot threshold being 15%, or other values, or they can be adjusted according to actual needs, etc. This application does not limit them.

[0067] For example, if the first blanking threshold is 85% and the second blanking threshold is 25%, and the blanking rate is determined to be 88% based on the ratio of the number of deleted invalid images to the batch processing threshold N, which is greater than the first blanking threshold, then it indicates that the current environment has significant interference, and a sensitivity downgrade instruction can be generated. If the blanking rate is determined to be 15% based on the ratio of the number of deleted invalid images to the batch processing threshold N, which is less than the second blanking threshold, then it indicates that the current animal activity is frequent, and a sensitivity upgrade instruction can be generated. If the blanking rate is determined to be 80% based on the ratio of the number of deleted invalid images to the batch processing threshold N, which is less than the first blanking threshold but greater than the second blanking threshold, then the current sensitivity can be maintained.

[0068] It should be noted that the above examples are merely illustrative and should not be construed as limiting the first no-shot threshold, second no-shot threshold, sensitivity adjustment command, etc. in the embodiments of this application.

[0069] Therefore, in this embodiment, if the false alarm rate is too high, it indicates that the current environmental interference is strong, such as wind blowing leaves or hot air currents causing frequent false triggers. The computing unit generates a sensitivity downgrade instruction. If the false alarm rate is too low, it indicates that the current monitoring area is frequently active. The computing unit generates a sensitivity upgrade instruction. Thus, based on the relationship between the false alarm rate and the false alarm threshold, the degree of interference in the current monitoring environment and the animal activity can be accurately determined, and targeted sensitivity adjustment instructions can be generated to achieve dynamic optimization of the trigger threshold of the front-end thermal infrared sensor.

[0070] Optionally, the computing unit can be used to detect the number of currently unprocessed captured images after generating the sensitivity adjustment instruction, and send a power-off request to the microcontroller if the number of unprocessed captured images is less than the batch processing threshold.

[0071] During the processing of the previous batch of captured images by the computing unit, which has reached the batch processing threshold, the acquisition unit may continue to acquire and store new images due to triggering by the thermal infrared sensor. Therefore, after the computing unit completes the processing of the previous batch of captured images and generates sensitivity adjustment instructions, the cumulative amount of unprocessed captured images in the current storage area can be further determined. If the number of unprocessed captured images is less than the batch processing threshold, it indicates that further processing is not necessary at this time. The computing unit can send a power-off request to the microcontroller, so that the microcontroller can control the computing unit to enter a power-off sleep state until it is woken up again after the number of subsequent microprocessed captured images reaches the batch processing threshold, thereby reducing the energy consumption of the computing unit during idle operation. Alternatively, if the number of unprocessed captured images has met the batch processing threshold, the computing unit can start a new round of batch processing, including target monitoring, empty shot rate calculation, and sensitivity adjustment instruction generation, thereby avoiding the energy waste and time loss caused by frequent power-offs and wake-ups, and effectively ensuring the continuity of image processing.

[0072] Optionally, the microcontroller can be used to look up the corresponding target cumulative variable in the mapping data table based on the blanking rate when a sensitivity adjustment command is received, and process the reference voltage trigger threshold, ambient temperature, preset compensation coefficient and target cumulative variable to update the voltage trigger threshold.

[0073] The mapping data table stores the mapping relationship between different no-shot rate ranges and corresponding cumulative variables. For example, a no-shot rate of 80%-85% corresponds to a cumulative variable of +0.1V, a no-shot rate of 15%-20% corresponds to a cumulative variable of -0.2V, and so on. When the microcontroller receives a sensitivity adjustment command, it can traverse the mapping data table according to the no-shot rate carried in the sensitivity adjustment command to determine the target cumulative variable corresponding to that no-shot rate. Then, it combines the reference voltage trigger threshold, ambient temperature value, preset compensation coefficient, and target cumulative variable to calculate and update the voltage trigger threshold, realizing the dynamic correction and update of the voltage trigger threshold.

[0074] In addition, the reference voltage trigger threshold can be the minimum voltage value at which the thermal infrared sensor can sensitively detect the target animal under ideal conditions of 25°C standard room temperature, no wind, and no interference. This threshold can correspond to the model and specifications of the thermal infrared sensor. The preset compensation coefficient can be a temperature compensation parameter matched to the model of the thermal infrared sensor. It can be calibrated by the sensor at the factory or calibrated according to the actual monitoring scenario. This application does not limit this.

[0075] Alternatively, the current ambient temperature value can be subtracted from 25°C to obtain the ambient temperature difference. Then, the ambient temperature difference can be multiplied by a preset compensation coefficient to obtain the temperature compensation value. The difference between the reference voltage trigger threshold and this temperature compensation value is then summed with the target cumulative variable, and the result is used as the updated voltage trigger threshold.

[0076] Optionally, the microcontroller can also be used to perform boundary verification on the updated voltage trigger threshold after completing the update calculation. If the voltage trigger threshold is less than the minimum boundary threshold, the minimum boundary threshold is determined as the voltage trigger threshold. If the voltage trigger threshold is greater than the maximum boundary threshold, the maximum boundary threshold is determined as the voltage trigger threshold.

[0077] Among them, the minimum boundary threshold is the minimum allowable value of the voltage trigger threshold, and the maximum boundary threshold is the maximum allowable value of the voltage trigger threshold. Both can be preset according to the performance parameters of the thermal infrared sensor, the actual needs of the field monitoring scenario, etc., and this application does not limit them.

[0078] Therefore, in this embodiment, the boundary verification described above can effectively prevent extreme values ​​of the voltage trigger threshold caused by factors such as abnormal fluctuations in the no-shot rate and temperature compensation deviations. For example, if the voltage trigger threshold is too high, the analog voltage value output by the thermal infrared sensor may be difficult to reach the threshold, and the microcontroller will not trigger the acquisition unit to start, resulting in missed monitoring of the target animal. If the voltage trigger threshold is too low, minor interference in the environment will cause the sensor to frequently trigger falsely, causing the acquisition unit to work continuously, which wastes system energy and generates a large amount of invalid no-shot data. Therefore, limiting the voltage trigger threshold to the range between the minimum and maximum boundary thresholds can effectively ensure that the system is always in a stable monitoring state, avoid system failure caused by threshold abnormalities, and thus improve the overall monitoring reliability.

[0079] Optionally, the microcontroller can also be used to perform a power-off cold start on the computing unit when the computing unit enters the power-on working state and no response signal is received from the computing unit within a preset time.

[0080] Understandably, after the microcontroller switches the computing unit from a power-off sleep state to a power-on working state, it may activate a built-in timing unit, such as a timer, and preset a response timeout, such as 10 seconds, 30 seconds, 60 seconds, etc. During this preset time, the microcontroller will continuously listen for the startup response signal returned by the computing unit through the UART interface. If the microcontroller does not receive a response signal from the computing unit within the preset time, it can determine that the computing unit has a startup abnormality, such as system freeze, initialization failure, or communication failure. In this case, the microcontroller can cut off the power supply to the computing unit, maintain the power-off state for the preset time, and then restore power to the computing unit, completing the cold start process for the computing unit.

[0081] Therefore, in this embodiment of the application, by performing a power-off cold start on the computing unit when it fails to respond after a timeout, the problem of occasional startup failure of the computing unit in complex field environments can be effectively solved. It can be automatically reset without manual intervention, avoiding interruption of the image processing flow due to the computing unit being stuck, ensuring the continuity of the system in batch image processing and feedback sensitivity adjustment instructions, and further improving the stability and robustness of the system in unattended field scenarios.

[0082] Optionally, the field monitoring system may also include a power management unit, wherein the power management unit includes a first switching device and a second switching device.

[0083] The first switching device can be turned on when the microcontroller outputs a high-level signal, so that the acquisition unit enters a powered-on working state, and turned off when the microcontroller outputs a low-level signal, so that the acquisition unit enters a power-off sleep state. The second switching device can be turned on when the microcontroller outputs a high-level signal, so that the computing unit enters a powered-on working state, and turned off when the microcontroller outputs a low-level signal, so that the computing unit enters a power-off sleep state.

[0084] The first and second switching devices can be N-channel MOSFETs, the first terminal of the microcontroller can be a first GPIO interface, and the second terminal of the microcontroller can be a second GPIO interface.

[0085] The control terminal of the first switching device is electrically connected to the first GPIO interface of the microcontroller, the input terminal is connected to the system power supply terminal (i.e., the battery), and the output terminal is connected to the power input terminal of the acquisition unit. When the microcontroller outputs a high-level signal through the first GPIO interface, the first switching device is turned on, the system power supply supplies power to the acquisition unit, and the acquisition unit switches from a power-off sleep state to a power-on working state. When the first GPIO interface of the microcontroller outputs a low-level signal, the first switching device is turned off, the power supply link of the acquisition unit is cut off, and it returns to the power-off sleep state.

[0086] The control terminal of the second switching device is electrically connected to the second GPIO interface of the microcontroller, the input terminal is connected to the system power supply terminal (i.e., the battery), and the output terminal is connected to the power input terminal of the computing unit. When the microcontroller outputs a high-level signal through the second GPIO interface, the second switching device is turned on, the system power supply supplies power to the computing unit, and the computing unit switches from the power-off sleep state to the power-on working state. When the second GPIO interface of the microcontroller outputs a low-level signal, the second switching device is turned off, the power supply link of the computing unit is cut off, and it returns to the power-off sleep state.

[0087] Therefore, in this embodiment, the power management unit independently controls the power supply to the acquisition unit and the computing unit through independent first and second switching devices. This enables both to be powered on demand and to sleep independently, minimizing energy consumption in non-working states to extend battery life in the field. Furthermore, the decoupling of power supply between modules avoids interference, improving control accuracy and system stability. At the same time, the fast response and low loss characteristics of MOSFET devices, combined with standardized interface design, further simplify the hardware architecture and enhance system scalability and compatibility.

[0088] Understandably, the microcontroller can serve as the primary control unit, the acquisition unit as the secondary acquisition unit, and the computing unit as the tertiary computing unit. The primary control unit communicates with the secondary acquisition unit and the tertiary computing unit via a UART serial communication link to achieve bidirectional transmission of instructions and data. The primary control unit outputs high and low level signals to the corresponding switching devices of the secondary acquisition unit and the tertiary computing unit through two independent GPIO interfaces to realize the power supply start and stop control of the secondary acquisition unit and the tertiary computing unit.

[0089] The field monitoring system based on image accumulation provided in this application is applicable to any field detection scenario. The following section combines... Figure 2 A brief description of the field monitoring system based on image accumulation provided in this application is given.

[0090] The primary control unit, also known as the microcontroller, can be the XIAO C3 micro embedded development board. The secondary acquisition unit can be the XIAO S3 as the control core of the camera. The tertiary computing unit can be the fourth-generation single-board computer Raspberry Pi 4B as the edge computing device.

[0091] Understandably, the primary control unit can receive the analog voltage value output by the thermal infrared PIR sensor, the current ambient temperature value collected by the temperature sensor, and the timestamp from the RTC clock module, which can be a DS3231 model. Furthermore, the primary control unit can control the operating states of MOSFET switch 1 and MOSFET switch 2 via GPIO, communicate with the secondary acquisition unit and the tertiary computing unit via UART / GPIO, and the power module, i.e., the battery, can power the primary control unit.

[0092] In addition, the primary control unit can control the MOSFET switch 1 to turn on and off through the output level of the first GPIO, thereby controlling the switching between the power-on operation and power-off sleep state of the secondary acquisition unit. The power input comes from the battery.

[0093] After the secondary acquisition unit is powered on, it starts the camera to capture images and writes them to the shared storage module SD card. When the number of captured images reaches the batch processing threshold, it can send a full-load cleaning request or a power-off request to the primary main control unit via UART / GPIO. Upon receiving a full-load cleaning request, the primary main control unit can output a high level via the second GPIO to control MOSFET switch 2 to conduct, thus powering on the secondary acquisition unit, which receives power from the battery. Upon receiving a power-off request from the secondary acquisition unit, the primary main control unit can output a low level via the first GPIO to put the secondary acquisition unit into a power-off sleep state.

[0094] After the third-level acquisition unit is powered on and awakened, it can read captured images from the shared storage module SD card, perform operations such as target detection and empty shot rate calculation, and send sensitivity adjustment commands or power-off requests to the first-level main control unit via UART / GPIO after processing. Upon receiving a power-off request from the third-level computing unit, the first-level main control unit can control the second GPIO to output a low level, causing the third-level computing unit to enter a power-off sleep state.

[0095] It should be noted that the above examples are merely illustrative and should not be construed as limiting the microcontroller, acquisition unit, computing unit, RTC clock module, etc. in the embodiments of this application.

[0096] In this embodiment, the field monitoring system based on image accumulation includes a microcontroller, an acquisition unit, a computing unit, and a sensing unit. The microcontroller switches the acquisition unit between a power-off sleep state and a power-on working state based on the relationship between the analog voltage value output by the thermal infrared sensor and the voltage trigger threshold. When the acquisition unit is in the power-on working state, it captures and stores images locally. When the number of captured images reaches the batch processing threshold, it sends a full-load cleaning request to the microcontroller. Upon receiving the full-load cleaning request, the microcontroller controls the computing unit to switch from the power-off sleep state to the power-on working state. When the computing unit is in the power-on working state, it performs batch processing on the stored captured images, generates a sensitivity adjustment command based on the no-shot rate, and sends it to the microcontroller so that the microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the current ambient temperature value. Therefore, by working together with the sensing unit, microcontroller, acquisition unit and computing unit, and combining image accumulation and batch processing mechanisms, the energy consumption of the computing unit can be effectively reduced during start-up and shutdown, and the endurance in the field can be greatly extended. At the same time, by dynamically adjusting the voltage trigger threshold of the acquisition unit in combination with the ambient temperature and the no-shot rate, false triggers caused by environmental interference can be effectively reduced, and the missed monitoring of target animals can be avoided. This significantly improves the reliability of ecological monitoring in unattended field scenarios and fully meets the diverse field monitoring needs.

[0097] like Figure 3 As shown, this field monitoring method based on image accumulation may include the following steps:

[0098] Step 301: The microcontroller switches the power-off sleep state and the power-on working state of the acquisition unit according to the relationship between the analog voltage value and the voltage trigger threshold. The analog voltage value is the analog voltage value output by the thermal infrared sensor after converting the infrared radiation changes it senses.

[0099] When the field monitoring system is first started, the voltage trigger threshold can be a pre-set reference voltage trigger threshold, or it can be pre-compensated based on the current ambient temperature. This application does not limit this.

[0100] Step 302: When the acquisition unit is powered on, it performs image capture and local storage, and when the number of captured images reaches the batch processing threshold, it sends a full-load cleaning request to the microcontroller.

[0101] The batch processing threshold can be a pre-set value, such as 20, 35, 40, etc., or it can be adjusted according to the actual needs such as the model and specifications of the acquisition unit, the frequency of animal activity in the wild environment, and the processing performance of the computing unit. This application does not limit this.

[0102] Optionally, the acquisition unit can also send an acquisition end command to the microcontroller if the number of image captures has not reached the batch processing threshold. Upon receiving the acquisition end command, the microcontroller can control the acquisition unit to switch from a power-on working state to a power-off sleep state.

[0103] Step 303: When the microcontroller receives a full-load cleaning request, it controls the computing unit to switch from a power-off sleep state to a power-on working state.

[0104] Step 304: When the computing unit is powered on, it performs batch processing on the stored captured images, generates a sensitivity adjustment command based on the no-shot rate of the captured images in this batch processing, and sends it to the microcontroller.

[0105] Optionally, the computing unit can perform target detection processing on the stored captured images to identify target images containing animal data and invalid images that do not contain animal data. Based on the number of invalid images and the batch processing threshold, the empty capture rate is determined, and invalid images are deleted.

[0106] Optionally, the sensitivity adjustment command may include a sensitivity downgrade command and a sensitivity upgrade command. The computing unit may generate a sensitivity downgrade command when the blank rate is greater than a first blank rate threshold, and generate a sensitivity upgrade command when the blank rate is less than a second blank rate threshold.

[0107] Optionally, after generating sensitivity adjustment instructions, the computing unit can also detect the number of currently unprocessed captured images. If the number of unprocessed captured images is less than the batch processing threshold, it indicates that further processing is not necessary at this time. The unit can then send a power-off request to the microcontroller, allowing the microcontroller to control the computing unit into a power-off sleep state until the number of captured images processed reaches the batch processing threshold, thus reducing the energy consumption of the computing unit during idle operation. Alternatively, if the number of unprocessed captured images already meets the batch processing threshold, the computing unit can initiate a new round of batch processing, including target monitoring, empty-shot rate calculation, and sensitivity adjustment instruction generation. This avoids the energy waste and time loss caused by frequent power-offs and wake-ups, effectively ensuring the continuity of image processing.

[0108] Optionally, the microcontroller can also perform a power-off cold start on the computing unit if it enters the power-on working state and does not receive a response signal from the computing unit within a preset time.

[0109] Optionally, the microcontroller can output a high-level signal to turn on the first switching device, so that the acquisition unit enters the power-on working state, and output a low-level signal to turn off the first switching device, so that the acquisition unit enters the power-off sleep state.

[0110] Optionally, the microcontroller can also output a high-level signal to turn on the second switching device, so that the computing unit enters the power-on working state, and output a low-level signal to turn off the second switching device, so that the computing unit enters the power-off sleep state.

[0111] Step 305: The microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the current ambient temperature value, wherein the current ambient temperature value is acquired by the temperature sensor.

[0112] Optionally, upon receiving a sensitivity adjustment command, the microcontroller can look up the corresponding target cumulative variable in the mapping data table based on the blank rate included in the sensitivity adjustment command, and process the reference voltage trigger threshold, ambient temperature, preset compensation coefficient, and target cumulative variable to update the voltage trigger threshold.

[0113] Optionally, after completing the update calculation of the voltage trigger threshold, the microcontroller can also perform boundary verification on the updated voltage trigger threshold. If the voltage trigger threshold is less than the minimum boundary threshold, the minimum boundary threshold is determined as the voltage trigger threshold. If the voltage trigger threshold is greater than the maximum boundary threshold, the maximum boundary threshold is determined as the voltage trigger threshold.

[0114] The following is combined with Figure 4The field monitoring process based on image accumulation provided in the embodiments of this application will be described.

[0115] For ease of description, the microcontroller can be abbreviated as C3, the acquisition unit as S3, and the computing unit is a Raspberry Pi.

[0116] Understandably, when conducting field monitoring based on accumulated image data, environmental perception and dynamic baseline setting can be performed first. At this point, C3 enters sleep mode for monitoring. Subsequently, C3 can read environmental data such as the current ambient temperature value and trigger the reference voltage threshold based on this ambient temperature value. Physical compensation, also known as pre-compensation, is performed to obtain the voltage trigger threshold. C3 then reads the PIR analog voltage value. If the PIR analog voltage value is greater than the current voltage trigger threshold, it performs graded wake-up and image accumulation, waking up S3. S3 captures images and stores them in shared storage, with a counter N=N+1. After the capture ends, it checks if the accumulated number N is greater than or equal to the batch threshold. If not, S3 notifies C3 that acquisition is complete, and C3 cuts off the power to S3. If yes, S3 sends a full-load cleaning request to C3, and C3 wakes up the Raspberry Pi to perform batch calculation and cleaning. The Raspberry Pi reads N images in batches and uses AI inference to perform image cleaning based on the inference results, deleting invalid empty images. Then, the Raspberry Pi calculates the empty shot rate and performs closed-loop feedback adjustment. If the empty shot rate is greater than 80% of the first empty shot threshold, a sensitivity downgrade command can be generated to increase the voltage trigger threshold. If the empty shot rate is less than 20% of the second empty shot threshold, a sensitivity upgrade command can be generated to decrease the voltage trigger threshold. Afterwards, the Raspberry Pi can send a sensitivity adjustment command to C3 via UART to request a power-off. Upon receiving the sensitivity adjustment command, C3 adjusts the voltage trigger threshold.

[0117] The following is combined with Figure 5 The dynamic calculation and trigger control process of the voltage trigger threshold inside the primary main control unit C3 provided in the embodiments of this application is described.

[0118] The primary control unit's input section is divided into two main data sources: input 1 is ambient temperature data. The real-time ambient temperature value from the temperature sensor serves as the input for physical layer compensation, correcting for temperature drift. Input 2 is the AI ​​no-shot rate feedback. The image batch processing results from the Raspberry Pi, a three-tier computing unit, are used as input for the data layer feedback to perform dynamic noise suppression. Then, the correction values ​​from the above two layers can be compared with the initial baseline threshold. The final execution threshold is obtained by summing the results. The comparator receives the final execution threshold. The PIR sensor's real-time analog signal and the PIR sensor's analog signal are compared numerically to determine whether the PIR signal is greater than 1. If the PIR real-time signal is greater than The comparator generates a wake-up trigger signal and sends it to the secondary acquisition unit S3, triggering the secondary acquisition unit to switch from power-off sleep mode to power-on working mode.

[0119] Therefore, the primary control unit uses a dual-dimensional correction mechanism based on ambient temperature and idle rate. This mechanism not only compensates for temperature drift errors in the sensors by adjusting temperature data but also dynamically suppresses environmental noise by incorporating idle rate feedback. This ensures that the final execution threshold can be adapted to the current scenario, effectively avoiding false triggering and missed monitoring issues under a single threshold. Furthermore, the closed-loop feedback of acquisition, processing, and threshold adjustment adapts to complex field environments. Through precise signal comparison using a comparator, the acquisition unit is only awakened when a real target is triggered. This improves the accuracy of monitoring triggering and the system's adaptability while reducing ineffective energy consumption, fully meeting the monitoring needs of long-term unattended operation in the field.

[0120] In this embodiment, the microcontroller can switch between the power-off sleep state and the power-on working state of the acquisition unit based on the relationship between the analog voltage value and the voltage trigger threshold. When the acquisition unit is in the power-on working state, it performs image capture and local storage. When the number of captured images reaches the batch processing threshold, it sends a full-load cleaning request to the microcontroller, so that the microcontroller controls the computing unit to switch from the power-off sleep state to the power-on working state. When the computing unit is in the power-on working state, it performs batch processing on the stored captured images, generates a sensitivity adjustment command based on the no-shot rate, and sends it to the microcontroller, so that the microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the current ambient temperature value. Therefore, by working together with the sensing unit, microcontroller, acquisition unit and computing unit, and combining image accumulation and batch processing mechanisms, the energy consumption of the computing unit can be effectively reduced during start-up and shutdown, and the endurance in the field can be greatly extended. At the same time, by dynamically adjusting the voltage trigger threshold of the acquisition unit in combination with the ambient temperature and the no-shot rate, false triggers caused by environmental interference can be effectively reduced, and the missed monitoring of target animals can be avoided. This significantly improves the reliability of ecological monitoring in unattended field scenarios and fully meets the diverse field monitoring needs.

[0121] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification, provided that such combination does not contain contradictions.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A field monitoring system based on image accumulation, characterized in that, include: Microcontroller, acquisition unit, computing unit and sensing unit; The sensing unit includes a thermal infrared sensor and a temperature sensor. The thermal infrared sensor is used to convert the sensed changes in infrared radiation into an analog voltage value, and the temperature sensor is used to collect the current ambient temperature value. The microcontroller is used to acquire the analog voltage value output by the thermal infrared sensor, and to switch the power-off sleep state and the power-on working state of the acquisition unit according to the relationship between the analog voltage value and the voltage trigger threshold. The acquisition unit is used to capture and store images locally when it is powered on, and to send a full-load cleaning request to the microcontroller when the number of captured images reaches the batch processing threshold. The microcontroller is also used to control the computing unit to switch from a power-off sleep state to a power-on working state when it receives the full-load cleaning request. The computing unit is used to perform batch processing on the stored captured images when the power is on. It generates a sensitivity adjustment command based on the no-shot rate of the captured images in this batch processing and sends it to the microcontroller so that the microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the current ambient temperature value. The microcontroller, upon receiving the sensitivity adjustment instruction, searches the mapping data table based on the no-shot rate to determine the corresponding target cumulative variable, and processes the reference voltage trigger threshold, the ambient temperature value, the preset compensation coefficient, and the target cumulative variable to update the voltage trigger threshold. The sensitivity adjustment commands include sensitivity downgrade commands and sensitivity upgrade commands; The computing unit is used to generate a sensitivity downgrade instruction when the empty shot rate is greater than a first empty shot threshold, and to generate a sensitivity upgrade instruction when the empty shot rate is less than a second empty shot threshold. The computing unit is used to detect the number of currently unprocessed captured images after generating the sensitivity adjustment instruction, and to send a power-off request to the microcontroller if the number of unprocessed captured images is less than the batch processing threshold.

2. The field monitoring system as described in claim 1, characterized in that, The acquisition unit is used to send an acquisition end command to the microcontroller when the number of image captures has not reached the batch processing threshold. The microcontroller is used to control the acquisition unit to switch from a powered-on working state to a powered-off sleep state when it receives the acquisition end command.

3. The field monitoring system as described in claim 1, characterized in that, The computing unit is used to perform target detection processing on the stored captured images to identify target images containing animal data and invalid images that do not contain animal data. The unit determines the no-shot rate based on the number of invalid images and the batch processing threshold, and deletes the invalid images.

4. The field monitoring system as described in claim 1, characterized in that, The microcontroller is configured to determine the minimum boundary threshold as the voltage trigger threshold when the voltage trigger threshold is less than the minimum boundary threshold, and to determine the maximum boundary threshold as the voltage trigger threshold when the voltage trigger threshold is greater than the maximum boundary threshold.

5. The field monitoring system as described in claim 1, characterized in that, The microcontroller is used to perform a power-off cold start on the computing unit when it is controlled to enter the power-on working state and no response signal is received from the computing unit within a preset time.

6. The field monitoring system as described in claim 1, characterized in that, The field monitoring system includes a power management unit; The power management unit includes a first switching device and a second switching device; The first switching device is used to turn on when the microcontroller outputs a high-level signal at its first terminal, so that the acquisition unit enters a powered-on working state, and to turn off when the microcontroller outputs a low-level signal, so that the acquisition unit enters a powered-off sleep state. The second switching device is used to turn on when the microcontroller outputs a high-level signal at its second terminal, so that the computing unit enters a powered-on working state, and to turn off when the microcontroller outputs a low-level signal, so that the computing unit enters a power-off sleep state.

7. A field monitoring method based on image accumulation, characterized in that, The field monitoring system based on image accumulation as described in any one of claims 1-6 includes: The microcontroller switches between the power-off sleep state and the power-on working state of the acquisition unit based on the relationship between the analog voltage value and the voltage trigger threshold. The analog voltage value is the analog voltage value output by the thermal infrared sensor after converting the infrared radiation changes it senses. When the acquisition unit is powered on, it captures and stores images locally, and sends a full-load cleaning request to the microcontroller when the number of captured images reaches the batch processing threshold. Upon receiving the full-load cleaning request, the microcontroller controls the computing unit to switch from a power-off sleep state to a power-on working state. When the computing unit is powered on, it performs batch processing on the stored captured images, generates a sensitivity adjustment command based on the no-shot rate of the captured images in this batch processing, and sends it to the microcontroller. The microcontroller adjusts the voltage trigger threshold according to the sensitivity adjustment command and the current ambient temperature value; The microcontroller, upon receiving the sensitivity adjustment instruction, searches the mapping data table based on the no-shot rate to determine the corresponding target cumulative variable, and processes the reference voltage trigger threshold, the ambient temperature value, the preset compensation coefficient, and the target cumulative variable to update the voltage trigger threshold. The sensitivity adjustment commands include sensitivity downgrade commands and sensitivity upgrade commands; The computing unit is used to generate a sensitivity downgrade instruction when the empty shot rate is greater than a first empty shot threshold, and to generate a sensitivity upgrade instruction when the empty shot rate is less than a second empty shot threshold. The computing unit is used to detect the number of currently unprocessed captured images after generating the sensitivity adjustment instruction, and to send a power-off request to the microcontroller if the number of unprocessed captured images is less than the batch processing threshold.

Images