Safety locking device, safety locking method and unmanned aerial vehicle laser obstacle clearing system

Abstract

This invention provides a safety interlocking device, a safety interlocking method, and a UAV laser obstacle removal system. The device includes: an acquisition unit for acquiring images of a laser-guided area; an identification unit for determining whether composite insulators exist in the image; and a closed-loop self-testing unit for driving a diffractive reference plate to alternately switch between a fully transparent first state and a second state projecting a standard feature map of composite insulators at a predetermined frequency to generate a self-test image. The acquisition unit is also used to acquire the self-test image. The identification unit is configured to, in self-test mode, determine whether the safety interlocking device is functioning correctly based on the matching relationship between the switching state of the diffractive reference plate and the image recognition result of the self-test image, thereby controlling the on / off state of the execution unit. This invention has the advantages of dedicated composite insulator identification, real-time closed-loop self-testing, and millisecond-level active interlocking, avoiding accidental damage to composite insulators during operation.

Description

Technical Field

[0001] This invention relates to the field of power transmission line obstacle clearing and power equipment safety protection technology, and in particular to a safety interlocking device, a safety interlocking method, and a drone laser obstacle clearing system. Background Technology

[0002] Using drones equipped with laser obstacle clearing devices to remove foreign objects (such as kite strings and tree branches) from high-voltage transmission lines has become a highly efficient and mainstream method. However, drones are prone to positional deviations due to wind disturbances in the air, which may cause the laser beam to deviate from its target and accidentally illuminate line equipment. The composite insulators on high-voltage transmission lines, whose skirts and core rods are made of polymer materials (such as silicone rubber and epoxy resin), are extremely sensitive to laser energy. Even short-term accidental irradiation can cause ablation and carbonization, leading to a permanent decline in insulation performance and causing major safety accidents. On the other hand, the metal components on the line (such as conductors and fittings) have a certain degree of tolerance to lasers and are not the main source of risk.

[0003] Currently, safety protection for this type of operation relies entirely on the pilot's visual judgment and manual emergency stop. However, the accuracy and reaction speed of human eyes in identifying composite insulators are limited, and they cannot respond to sudden shaking of drones within milliseconds, posing an extremely high risk of accidental injury.

[0004] Furthermore, even if composite insulators can be identified using existing general obstacle recognition algorithms, the reliability of the protection system itself cannot be verified in real time during actual operation. If problems such as image acquisition failure, recognition algorithm abnormality, or hardware link failure occur, the entire protection function will be completely ineffective, resulting in a fatal safety vulnerability and failing to meet the absolute safety requirements for live-line work on transmission lines. Summary of the Invention

[0005] The purpose of this invention is to provide a safety interlocking device, a safety interlocking method, and a UAV laser obstacle removal system. The safety interlocking device has the ability to identify composite insulators, perform real-time closed-loop self-testing, and provide millisecond-level active interlocking, fundamentally solving the problems of accidental damage to composite insulators and low system reliability in UAV laser obstacle removal operations.

[0006] In a first aspect, the present invention provides a safety interlocking device, which is installed on a drone laser obstacle removal device and includes a data acquisition unit, an identification unit, a self-testing module and an execution unit. The acquisition unit has its optical central axis parallel to the laser emission path of the laser in the UAV laser obstacle removal device, and is used to acquire images of the laser-pointing area. The identification unit is electrically connected to the acquisition unit and is used to determine whether a composite insulator exists in the image acquired by the acquisition unit. The self-test module includes a diffraction reference plate and a closed-loop self-test unit disposed in front of the acquisition unit. The closed-loop self-test unit is electrically connected to the diffraction reference plate and the identification unit, respectively. The closed-loop self-test unit is used to drive the diffraction reference plate to alternately switch between a first state and a second state at a predetermined frequency to generate a self-test image. The first state is a fully transparent state, and the second state is a state of projecting a standard feature map of a composite insulator. The acquisition unit is also used to acquire the self-test image. The execution unit is connected in series in the main power supply circuit of the laser, and its control terminal is controlled by the output signal of the recognition unit. The recognition unit is configured to, in self-test mode, determine whether the safety interlocking device is normal based on the matching relationship between the switching state of the diffractive reference plate and the image recognition result of the self-test image, and control the on / off state of the execution unit based on the judgment result of the self-test mode.

[0007] Furthermore, the identification unit is also configured to, in monitoring mode, determine whether the UAV laser obstacle clearing device has accidentally damaged the composite insulator based on the image recognition results of the on-site images collected by the acquisition unit, and control the on / off state of the execution unit based on the judgment results of the monitoring mode; The self-test mode is specifically as follows: The closed-loop self-test unit drives the diffractive reference plate to alternately switch between the first state and the second state at a predetermined frequency based on the reference signal to generate a self-test image. The acquisition unit synchronously acquires the self-test image; The identification unit performs a logical self-test based on the switching state of the diffractive reference plate and the identification result of the self-test image. If the self-test fails, the execution unit is driven to disconnect the main power supply circuit of the laser and report the fault. If the self-test succeeds, the unit enters the monitoring mode. The monitoring mode is specifically as follows: The closed-loop self-testing unit drives the diffractive reference plate to display the first state based on the self-test success signal; The acquisition unit acquires real-time images of the laser-pointing area in the actual work scenario; The identification unit analyzes the scene image. If it identifies a composite insulator within the laser hazard area of ​​the scene image, it drives the execution unit to disconnect the main power supply circuit of the laser to achieve active blocking. If it does not identify a composite insulator within the laser hazard area of ​​the scene image, it keeps the laser enabled.

[0008] Furthermore, in the self-test mode, the recognition unit performs logical self-test verification based on the switching state of the diffractive reference plate and the image recognition result of the self-test image, including: If the identification unit identifies composite insulator features in the self-test image when the diffractive reference plate is in the second projection state, and does not identify composite insulator features in the self-test image when the diffractive reference plate is in the first state, then the self-test is determined to be successful. If the identification unit fails to identify composite insulator features in the self-test image when the diffractive reference plate is in the second projection state, and / or identifies composite insulator features in the self-test image when the diffractive reference plate is in the first state, then the self-test is determined to have failed.

[0009] Furthermore, the self-test mode of the identification unit is activated when the safety interlocking device is powered on and initialized, during work breaks, or when a manual trigger command is received. The execution unit is configured to require manual intervention to reset after being disconnected. Before manual intervention, the execution unit maintains the disconnected state of the laser's main power supply circuit.

[0010] Furthermore, the closed-loop self-testing unit drives the diffractive reference plate to alternate between the first state and the second state at a frequency of not less than 5Hz; After the identification unit detects the presence of a composite insulator within the laser hazard area of ​​the on-site image, it sends a locking command to the execution unit within 10ms. Upon receiving the locking command, the execution unit immediately cuts off the main power supply circuit of the laser, and the safety locking device simultaneously records the image data and system status information at the locking moment. The safety interlocking device is configured to automatically trigger a new round of self-inspection process after each continuous operation reaches a preset time.

[0011] Furthermore, the recognition unit incorporates a dedicated image recognition algorithm for the shape, texture, and color features of composite insulators, and the dedicated image recognition algorithm employs convolutional neural network image recognition technology.

[0012] Furthermore, the acquisition unit, the identification unit, the execution unit, and the closed-loop self-testing unit are all integrated into the same housing to form an intelligent interlocking module; The diffractive reference plate is composed of a diffractive projection plate and an LED display module. The LED display module is electrically connected to the closed-loop self-test unit. The diffractive projection plate forms a standard feature map of a composite insulator. The execution unit includes a fast relay, the control terminal of which is controlled by the output signal of the identification unit; The acquisition unit is rigidly connected to the laser to maintain the parallelism between the optical central axis of the acquisition unit and the laser emission path of the laser.

[0013] In a second aspect, the present invention provides a safety interlocking method, implemented based on a safety interlocking device as described in any of the preceding claims, the safety interlocking method comprising: S100. The safety interlocking device is powered on and starts the self-test mode. The closed-loop self-test unit drives the diffractive reference plate to periodically switch between the first state and the second state at a predetermined frequency based on the reference signal to generate a self-test image. The acquisition unit synchronously acquires the self-test image. S200: The recognition unit judges the self-test image, performs logical self-test verification based on the switching state of the diffractive reference plate and the image recognition result of the self-test image. If the self-test fails, the execution unit is driven to disconnect the main power supply circuit of the laser and report the fault. If the self-test is successful, the monitoring mode in step S300 is entered. S300, Enter real-time monitoring mode. The closed-loop self-testing unit drives the diffractive reference plate to display in the first state based on the self-test success signal. The acquisition unit continuously acquires on-site images of the laser pointing area in the real operation scenario. S400, The identification unit performs real-time analysis on each frame of the scene image to determine whether there is a composite insulator within the laser hazard area of ​​the scene image; S500: If the composite insulator is detected, a locking command is issued to drive the execution unit to perform a locking action, cutting off the main power supply circuit of the laser to achieve active locking; if the composite insulator is not detected, the laser is kept in an enabled state. S600. During work breaks or when a manual instruction is received, a new round of self-check process is automatically or manually triggered, and steps S100 to S200 are repeated.

[0014] Furthermore, in the self-test mode, the recognition unit performs logical self-test verification based on the switching state of the diffractive reference plate and the image recognition result of the self-test image, including: If the identification unit identifies composite insulator features in the self-test image when the diffractive reference plate is in the second projection state, and does not identify composite insulator features in the self-test image when the diffractive reference plate is in the first state, then the self-test is determined to be successful. If the identification unit fails to identify composite insulator features in the self-test image when the diffractive reference plate is in the second projection state, and / or identifies composite insulator features in the self-test image when the diffractive reference plate is in the first state, then the self-test is determined to have failed.

[0015] Thirdly, the present invention provides a drone laser obstacle removal system, comprising: a drone laser obstacle removal device and a safety interlocking device as described in any of the above claims; The UAV laser obstacle removal device includes a laser, which includes a laser body and a laser lens. The laser lens is disposed in the laser emission path of the laser body. The acquisition unit, identification unit, execution unit, and closed-loop self-test unit in the safety interlocking device are integrated into the same housing to form an intelligent interlocking module. The intelligent interlocking module and the diffraction reference plate are both located above the laser body, and the diffraction reference plate is in close contact with the laser lens.

[0016] As can be seen from the above, the safety interlocking device, safety interlocking method, and UAV laser obstacle clearing system provided by this invention can effectively prevent composite insulators on high-voltage transmission lines from being accidentally damaged by the UAV laser obstacle clearing device by accurately acquiring images of the laser-pointing area through the acquisition unit and combining the recognition capability of the recognition unit, thus achieving precise protection. Simultaneously, the introduction of a self-testing module, using standard feature maps of composite insulators to simulate real threats, enables real-time verification of the working status of the image acquisition and recognition system itself. This simultaneously verifies the unobstructedness of the optical path, the normal operation of the acquisition unit, and the accuracy of the recognition algorithm of the recognition unit, overcoming the deficiency of lacking system reliability self-testing in existing technologies. Therefore, this device can ensure that even if the UAV experiences unexpected shaking or system components malfunction during laser obstacle clearing operations, the laser power supply can be cut off promptly and reliably, significantly improving the safety and reliability of the operation. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a safety locking device provided by the present invention.

[0018] Figure 2 This is a schematic diagram of the closed-loop self-test logic determination process of the safety interlocking device provided by the present invention.

[0019] Figure 3 This is a flowchart illustrating a safety interlocking method provided by the present invention.

[0020] In the diagram: 100, UAV laser obstacle removal system; 1, UAV laser obstacle removal device; 11, laser; 111, laser body; 112, laser lens; 2, safety interlocking device; 20, intelligent interlocking module; 21, acquisition unit; 22, identification unit; 23, self-testing module; 231, diffraction reference plate; 232, closed-loop self-testing unit; 24, execution unit. Detailed Implementation

[0021] The technical solutions of this invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some, not all, of the embodiments of this invention. The components of this invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0022] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0023] It should be noted that the method of this disclosure embodiment can be executed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method of this disclosure embodiment, and the multiple devices will interact with each other to complete the method described.

[0024] It should be noted that the above description describes some embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0025] Traditional drone-based laser obstacle removal systems pose a risk of damaging composite insulators sensitive to laser energy when clearing foreign objects from high-voltage power lines. This risk arises from wind disturbances causing the laser beam to deviate from the target. Current safety measures rely on manual visual judgment and emergency stops, which have limited accuracy and reaction speed, making them unable to respond promptly to sudden vibrations and resulting in a high risk of accidental damage. Furthermore, existing recognition algorithms cannot verify the reliability of the protection system itself in real time. Failures in image acquisition, recognition algorithms, or hardware links can lead to the failure of the protection function, creating a fatal security vulnerability.

[0026] To resolve the above issues, please refer to [link / reference]. Figure 1 and Figure 2 The present invention proposes a safety locking device 2, which is installed on the UAV laser obstacle clearing device 1 and includes a data acquisition unit 21, an identification unit 22, a self-test module 23 and an execution unit 24. The acquisition unit 21 has its optical central axis parallel to the laser emission path of the laser 11 in the UAV laser obstacle clearing device 1, and is used to acquire images of the laser-pointing area; The identification unit 22 is electrically connected to the acquisition unit 21 and is used to determine whether there is a composite insulator in the image acquired by the acquisition unit 21. The self-test module 23 includes a diffractive reference plate 231 and a closed-loop self-test unit 232 disposed in front of the acquisition unit 21. The closed-loop self-test unit 232 is electrically connected to the diffractive reference plate 231 and the identification unit 22, respectively. The closed-loop self-test unit 232 is used to drive the diffractive reference plate 231 to alternately switch between a first state and a second state at a predetermined frequency to generate a self-test image. The first state is a fully transparent state, and the second state is a state of projecting a standard feature map of a composite insulator. The acquisition unit 21 is also used to acquire the self-test image to perform self-test verification on the acquisition unit 21 and the identification unit 22. The execution unit 24 is connected in series in the main power supply circuit of the laser 11, and its control terminal is controlled by the output signal of the recognition unit 22. The recognition unit 22 is configured to, in self-test mode, determine whether the safety locking device 2 is normal based on the matching relationship between the switching state of the diffractive reference plate 231 and the image recognition result of the self-test image, and control the on / off state of the execution unit 24 based on the judgment result of the self-test mode.

[0027] In this embodiment, the safety interlocking device 2 is installed on the UAV laser obstacle clearing device 1 to prevent the UAV laser obstacle clearing device 1 from accidentally damaging the composite insulator. To better describe the overall relationship between the safety interlocking device 2 and the UAV laser obstacle clearing device 1, the two are referred to as the UAV laser obstacle clearing system 100, or simply the system.

[0028] The optical central axis of the acquisition unit 21 is precisely calibrated to ensure it is parallel to the laser emission path of the laser 11. The main function of this unit is to acquire image information of the laser-pointing area in real time, providing visual data for subsequent recognition and judgment. Furthermore, the acquisition unit 21 includes an industrial camera or a high-resolution camera.

[0029] The identification unit 22 is used to receive image data from the acquisition unit 21, and to analyze the image content using a specific algorithm to identify whether there is a composite insulator in the image, and output the corresponding identification result signal.

[0030] The self-test module 23 is used to verify the working status of the safety interlocking device 2 itself. In self-test mode, this module performs periodic or on-demand self-tests on the functions of the acquisition unit 21 and the recognition unit 22 by simulating the appearance and disappearance of the composite insulator and combining the image recognition results.

[0031] Specifically, the module includes a diffractive reference plate 231 and a closed-loop self-testing unit 232.

[0032] The diffraction reference plate 231 is a key component in the self-test module 23. It is positioned in front of the acquisition unit 21 and configured to switch between a first state and a second state at high speed under the control of the closed-loop self-test unit 232, i.e., it can display in two states: The first state is a fully transparent state, in which the diffractive reference plate 231 is fully transparent, does not block or change the light passing through it, and allows the acquisition unit 21 to directly acquire images of the external environment, so that the acquisition unit 21 can acquire images of the real scene in front of it. The second state is the projection of a standard feature map of a composite insulator. The diffractive reference plate 231 projects a preset standard feature map of a composite insulator to simulate the presence of a composite insulator during self-testing. This feature map is used to provide a clear image signal of a composite insulator to the acquisition unit 21 during self-testing to test the detection capability of the identification unit 22.

[0033] The closed-loop self-test unit 232 is the control part of the self-test module 23. It is responsible for driving the diffractive reference plate 231 to alternately switch between a first state and a second state to generate a self-test image. The self-test image includes the image of the diffractive reference plate 231 when it is in a fully transparent state and the image of the diffractive reference plate 231 when it is in the state of projecting the standard feature map of the composite insulator. At the same time, this unit also works in conjunction with the recognition unit 22, sending the first state image and second state image switching signal to the recognition unit 22 to complete the control and verification of the entire self-test process.

[0034] The execution unit 24 is used to control the on / off of the main power supply circuit of the laser 11. Its control terminal receives the output signal from the identification unit 22. When the identification unit 22 determines that locking is required, it will send a locking command to the execution unit 24. When the execution unit 24 receives the locking command, it can quickly cut off the power supply of the laser 11, thereby preventing laser emission and realizing the safe locking of the laser 11.

[0035] Furthermore, the execution unit 24 includes a fast relay. A fast relay is an electromagnetic relay with a fast response speed and reliable operation, capable of closing or opening its contacts in a very short time. Applying it to the execution unit 24 ensures that when the identification unit 22 detects a dangerous situation, it can cut off the main power supply circuit of the laser 11 at a speed of milliseconds or even faster, thereby achieving rapid locking and minimizing the risk of accidental injury.

[0036] In self-test mode, the recognition unit 22 is configured to compare the recognition result of the self-test image acquired by the acquisition unit 21 with the switching state of the diffraction reference plate 231: When the closed-loop self-test unit 232 drives the diffraction reference plate 231 to display the second state, the identification unit 22 should be able to identify the composite insulator features from the self-test image. When the closed-loop self-test unit 232 drives the diffraction reference plate 231 to display the first state, the identification unit 22 should not identify the composite insulator characteristics.

[0037] By determining this matching relationship, the unit can assess whether the entire safety interlocking device 2 (including the image acquisition and recognition section) is functioning properly. Based on this determination, the recognition unit 22 controls the on / off state of the execution unit 24, for example, cutting off the power to the laser 11 when the self-test fails, to ensure that the system does not malfunction in a fault state.

[0038] The safety interlocking device 2 provided in this embodiment of the invention accurately acquires images of the laser-pointing area through the acquisition unit 21, and combined with the recognition capability of the recognition unit 22, effectively prevents the UAV laser obstacle clearing device 1 from accidentally damaging composite insulators on high-voltage transmission lines, achieving precise protection. Simultaneously, the introduction of the self-testing module 23, using standard feature maps of composite insulators to simulate real threats, enables real-time verification of the working status of the image acquisition and recognition system itself. This simultaneously verifies the unobstructedness of the optical path, the normal operation of the acquisition unit 21, and the accuracy of the recognition algorithm of the recognition unit 22, overcoming the deficiency of system reliability self-testing in existing technologies. Therefore, this device ensures that even if the UAV experiences unexpected shaking or system components malfunction during laser obstacle clearing operations, the power supply to the laser 11 can be promptly and reliably cut off, significantly improving the safety and reliability of the operation.

[0039] In one embodiment, the identification unit 22 is further configured to, in monitoring mode, determine whether the UAV laser obstacle clearing device 1 has accidentally damaged the composite insulator based on the image recognition result of the on-site image collected by the acquisition unit 21, and control the on / off state of the execution unit 24 based on the judgment result of the monitoring mode. The self-test mode is specifically as follows: The closed-loop self-test unit 232 drives the diffractive reference plate 231 to alternately switch between the first state and the second state at a predetermined frequency based on the reference signal to generate a self-test image. The acquisition unit 21 synchronously acquires the self-test image; The identification unit 22 performs a logical self-test verification based on the switching state of the diffractive reference plate 231 and the identification result of the self-test image. If the self-test fails, it drives the execution unit 24 to disconnect the main power supply circuit of the laser 11 and report the fault. If the self-test succeeds, it enters the monitoring mode. The monitoring mode is specifically as follows: The closed-loop self-testing unit 232 drives the diffractive reference plate 231 to display the first state based on the self-test success signal; The acquisition unit 21 acquires real-time images of the laser-pointing area in the actual work scenario; The identification unit 22 analyzes the scene image. If it identifies a composite insulator in the laser hazard area of ​​the scene image, it drives the execution unit 24 to disconnect the main power supply circuit of the laser 11 to achieve active blocking. If it does not identify a composite insulator in the laser hazard area of ​​the scene image, it keeps the laser 11 enabled.

[0040] Specifically, the above self-check modes include: Based on a preset reference signal, the closed-loop self-test unit 232 precisely controls the diffractive reference plate 231 to alternately switch between a first state (fully transparent state) and a second state (projecting the standard feature map of the composite insulator) at a predetermined frequency. The self-test images then include images of the diffractive reference plate 231 in the first state and images of the diffractive reference plate 231 in the second state. This periodic switching provides the acquisition unit 21 and the recognition unit 22 with a series of known and controllable test image inputs. Setting the predetermined frequency, such as switching several times per second, ensures a thorough self-test is completed in a short time while avoiding unnecessary burden on system performance.

[0041] The switching actions of the acquisition unit 21 and the closed-loop self-test unit 232 are synchronized, and the self-test images displayed by the diffractive reference board 231 in different states are acquired in real time. This ensures that the acquired images can accurately reflect the current display state of the diffractive reference board 231, providing reliable input data for subsequent logic self-test verification. Specifically, the precise correspondence between image acquisition and state switching can be ensured by hardware triggering or software clock synchronization.

[0042] The identification unit 22 receives the self-test image acquired synchronously and, in conjunction with the current switching state of the diffraction reference plate 231 provided by the closed-loop self-test unit 232, performs a logic self-test verification to determine whether the identification unit can correctly identify or not identify the preset composite insulator features.

[0043] For example, when the diffractive reference plate 231 is in the second state of projecting the standard feature map of the composite insulator, the identification unit 22 should be able to successfully detect the composite insulator; however, when the diffractive reference plate 231 is in the first state of full transparency, the identification unit 22 should not detect the composite insulator. Through this logical comparison, it is possible to effectively detect whether there are false alarms or missed alarms in the identification unit 22 and the acquisition unit 21.

[0044] If the logic self-test verification fails, that is, the identification unit 22 fails to identify according to the expected result (for example, it fails to identify the composite insulator in the transparent state or fails to identify it when the composite insulator features are displayed), the execution unit 24 is immediately driven to disconnect the main power supply circuit of the laser 11 and report the fault information to the operator or control system. This prevents the system from operating under potential faults in the first instance and avoids accidental damage.

[0045] If the self-test verification is successful, it indicates that the identification unit 22 and the acquisition unit 21 are functioning normally, and the system will automatically enter the monitoring mode to prepare for the actual obstacle removal operation.

[0046] In one specific embodiment, after several consecutive self-test cycles have passed verification, the system indicator light turns green, indicating that the system is ready.

[0047] Therefore, through the above technical solution, the present invention can comprehensively and accurately evaluate the collaborative working state of the identification unit 22 and the acquisition unit 21. Specifically, by verifying the forward identification capability of the identification unit 22 when the diffractive reference plate 231 displays the standard feature diagram of the composite insulator, it ensures that it will not miss any reports in actual operation; at the same time, by verifying the reverse identification capability of the identification unit 22 when the diffractive reference plate 231 is in a fully transparent state, it ensures that it will not falsely report any reports in actual operation. This dual verification mechanism effectively avoids system failure caused by the decreased identification capability or false reporting tendency of the identification unit 22, thereby significantly improving the reliability and safety of the entire safety interlocking device 2. Once the self-test fails, the identification unit 22 can immediately drive the execution unit 24 to disconnect the main power supply circuit of the laser 11 and report the fault, thereby preventing the laser 11 from malfunctioning before potential danger occurs, maximizing the operational safety of the UAV laser obstacle clearing device 1 in complex working environments, and preventing accidental damage to the composite insulator.

[0048] Specifically, the above monitoring modes include: In monitoring mode, the closed-loop self-test unit 232 drives the diffractive reference plate 231 to display in a fully transparent first state based on the self-test success signal. In the transparent state, it does not interfere with the normal field of vision and ensures the smoothness of the obstacle clearing operation.

[0049] The acquisition unit 21 continuously and in real time acquires on-site images of the laser-pointing area in the actual work scenario, so as to continuously capture video streams or high frame rate image sequences of the area in front of the laser 11, ensuring that dynamic changes in the work environment are perceived in real time.

[0050] The identification unit 22 analyzes each frame of real-time acquired on-site image to determine whether there are composite insulators in the laser hazard area.

[0051] If a composite insulator is detected, a lockout command is sent to the execution unit 24, which will quickly cut off the main power supply circuit of the laser 11 to achieve active lockout. In this way, by adopting a hard safety logic of lockout upon detection, the reaction speed is far faster than that of a human, and the possibility of accidentally damaging the composite insulator is eliminated in principle, achieving absolute safety.

[0052] If no composite insulator is identified within the laser hazard area in the on-site image, the identification unit 22 keeps the laser 11 enabled, allowing the laser 11 to operate normally.

[0053] Through the above technical solution, this invention clearly defines the self-test mode and monitoring mode of the safety interlocking device 2, and specifies in detail the specific operation procedures and switching logic of the two modes. Specifically, in the self-test mode, the closed-loop self-test unit 232 drives the diffractive reference plate 231 to periodically switch the display of known feature maps, and the acquisition unit 21 synchronously acquires the self-test images. Then, the recognition unit 22 performs logical verification, which can comprehensively and reliably detect the functional integrity of the visual acquisition and recognition system. This rigorous self-test mechanism ensures that its core safety functions are reliable and effective before the system is put into actual operation, thereby avoiding the risk of misoperation caused by system failure. Once the self-test is successful, it seamlessly switches to the monitoring mode. The diffractive reference plate 231 is in a fully transparent first state, the acquisition unit 21 acquires images of the operation scene in real time, and the recognition unit 22 continuously analyzes the images to determine whether composite insulators are present. When a composite insulator is detected, the actuator 24 can be immediately driven to disconnect the main power supply circuit of the laser 11, achieving millisecond-level active interlocking. This greatly improves the safety of the UAV laser obstacle clearing device 1 in actual operation and effectively prevents accidental damage to composite insulators. This standardized operation and strict logical judgment enable the safety interlocking device 2 to provide continuous, efficient, and rapid-response operational protection while ensuring its own reliability.

[0054] In one embodiment, in the self-test mode, the recognition unit 22 performs logical self-test verification based on the switching state of the diffractive reference plate 231 and the image recognition result of the self-test image, including: If the identification unit 22 identifies composite insulator features in the self-test image when the diffractive reference plate 231 is in the second projection state, and does not identify composite insulator features in the self-test image when the diffractive reference plate 231 is in the first state, then the self-test is determined to be successful. If the identification unit 22 fails to identify composite insulator features in the self-test image when the diffractive reference plate 231 is in the second projection state, and / or identifies composite insulator features in the self-test image when the diffractive reference plate 231 is in the first state, then the self-test is determined to have failed.

[0055] The verification process includes the following two core judgment conditions: When the diffraction reference plate 231 is in the second state of projecting the standard feature map of composite insulators, that is, when it projects the preset standard feature map of composite insulators, it simulates the situation where composite insulators exist in a real scene. The identification unit 22 should be able to accurately detect and identify these preset composite insulator features from the self-test image acquired by the acquisition unit 21, and output the result as "composite insulator identified" to verify the positive identification capability of the identification unit when target features exist and ensure that its core identification function is normal.

[0056] When the diffractive reference plate 231 is in the first state of full transparency, that is, when it is in a fully transparent state, simulating the situation where there is no composite insulator in the real scene, the identification unit 22 should not identify the composite insulator feature from the self-test image acquired by the acquisition unit 21, and the output result is "no composite insulator identified", in order to verify the reverse identification capability of the identification unit when there is no target feature, that is, to avoid false alarms and ensure the accuracy of its identification.

[0057] The self-test is considered successful only when both of the above conditions are met: correct identification when features are present, and correct non-identification when features are absent (i.e., identification if an image is present, non-identification if no image is present). This indicates that the identification unit 22 and its front-end acquisition unit 21 are working properly together and have the ability to accurately identify targets and avoid false alarms in actual operations. It will then automatically enter monitoring mode, preparing for actual obstacle removal operations. This dual verification mechanism ensures that the identification unit 22 can correctly identify targets while avoiding false identification, thereby comprehensively assessing its functional integrity.

[0058] Conversely, when the diffractive reference plate 231 is in the second state of projecting the standard feature map of the composite insulator, the identification unit 22 fails to identify these features. This may mean that the identification ability of the identification unit 22 has decreased, the acquisition unit 21 is malfunctioning (e.g., image blurring, sensor obstruction), or there is a problem with communication between the two, causing the system to fail to detect the target correctly.

[0059] When the diffractive reference plate 231 is in its fully transparent first state, and should not exhibit composite insulator characteristics, the identification unit 22 incorrectly identifies composite insulator characteristics. This may indicate that the identification unit 22 has a tendency to generate false alarms, that the acquisition unit 21 has an anomaly (e.g., internal reflection, sensor noise), or that the self-test image processing logic has a defect, causing the system to generate false identifications when there is no target.

[0060] If any one or two of the above conditions occur, i.e., the identification unit 22 fails to identify when there is a feature, or incorrectly identifies when there is no feature, the self-test is deemed to have failed. This indicates that the core identification function of the safety interlocking device 2 is defective and cannot guarantee reliability in actual operation. The execution unit 24 should be immediately driven to disconnect the main power supply circuit of the laser 11 and report the fault information to the operator or control system to prevent the system from operating under potential fault conditions as soon as possible, thereby avoiding accidental injury.

[0061] Through the aforementioned clear logical judgment criteria, this invention constructs a rigorous and comprehensive self-testing verification mechanism. This mechanism not only verifies the identification capability of the identification unit 22 when a target is present, but also verifies its anti-false alarm capability when no target is present, thus forming a two-way, complementary verification process. This precise logical judgment can effectively avoid ambiguity or uncertainty in the self-testing process, ensuring the accuracy and reliability of the self-testing results. Once the identification unit experiences a failure such as identification omission or misidentification, the system can immediately determine the self-test failure and trigger corresponding safety interlocking measures, thereby significantly improving the safety of the UAV laser obstacle clearing device 1 operation and effectively preventing the risk of accidental damage to composite insulators.

[0062] In actual operation, the system may encounter new faults after the initial self-test is successful, or it may automatically reset after the interlock is triggered if the fault is not effectively handled. This may lead to repeated safety hazards, thereby affecting the continuous safety and reliability of the operation.

[0063] To address the aforementioned issues, the present invention further proposes a self-test mode for the identification unit 22, which is activated and executed when the safety interlocking device 2 is powered on and initialized, during work breaks, or when a manual trigger command is received.

[0064] Specifically, the self-test mode is no longer limited to the initial power-on phase, but extends to multiple critical nodes. For example, when the safety interlock device 2 is powered on and initialized, a comprehensive self-test will be automatically performed to ensure that all modules or units are in normal working condition at startup.

[0065] In addition, during the intervals of operation of the UAV laser obstacle clearing device 1, such as after completing an obstacle clearing task or reaching the preset operation time, a new round of self-inspection process can be automatically triggered to continuously verify the functional reliability of the acquisition unit 21 and the identification unit 22.

[0066] In addition, to address potential concerns from operators (pilots) or the need for pre-operation confirmation, the system also supports receiving manual trigger commands to initiate a self-test mode, such as sending commands via ground station control software or operating via physical buttons on the drone. This multi-timing trigger mechanism ensures continuous functional verification of the safety interlock device 2 throughout its operational lifecycle.

[0067] Therefore, by activating the self-test mode when the safety interlocking device 2 is powered on for initialization, during work breaks, or when a manual trigger command is received, the present invention ensures that the functions of the acquisition unit 21 and the identification unit 22 are always in a verified state throughout the entire work cycle, thereby improving the continuous reliability of the system. It can effectively solve the potential faults that may occur after the initial self-test is successful in UAV laser obstacle clearing operations, as well as the safety hazards that may be caused by automatic reset after interlocking triggering.

[0068] In one embodiment, the execution unit 24 is configured to require manual intervention to reset after disconnection. Before manual intervention, the execution unit 24 maintains the main power supply circuit of the laser 11 in a disconnected state.

[0069] When the execution unit 24 cuts off the main power supply circuit of the laser 11 due to the detection of a composite insulator or a self-test failure, it will remain in the disconnected state until it receives a clear manual reset command. After confirming that the danger has been eliminated or the fault has been repaired, if the pilot needs to adjust the drone's attitude until there are no composite insulators in the frame, specific operations, such as pressing the reset button, entering an authorization password, or sending a specific reset command, are required to restore the execution unit 24 to power. This forces the operator to investigate and confirm the cause of the lockout, preventing the system from automatically restoring power to the laser 11 without confirming safety, greatly improving operational safety and effectively preventing secondary injuries caused by system misjudgment or malfunction.

[0070] In one specific embodiment, the closed-loop self-test unit 232 drives the diffractive reference plate 231 to alternate between the first state and the second state at a frequency of not less than 5Hz.

[0071] This high-frequency switching ensures the speed and real-time nature of the self-test process, enabling the acquisition unit 21 to acquire self-test images under different states at short time intervals, and the recognition unit 22 to quickly perform logical self-test verification, promptly detect and report potential faults, thereby shortening the time the system is in an uncertain state.

[0072] Specifically, the closed-loop self-test unit 232 can precisely control the speed of display state switching by precisely controlling the drive circuit inside the diffractive reference board 231, for example, by using pulse width modulation (PWM) signals or digital switching signals.

[0073] To ensure swift action upon detection of a hazardous situation, in one embodiment, the identification unit 22 is configured to send a locking command to the execution unit 24 within 10 ms after identifying the presence of a composite insulator within the laser hazard area of ​​the scene image. Upon receiving the locking command, the execution unit 24 immediately cuts off the main power supply circuit of the laser 11, ensuring that the laser 11 stops operating.

[0074] This millisecond-level response speed enables millisecond-level active interlocking, greatly reducing the risk of laser damage to composite insulators.

[0075] In addition, to facilitate post-event analysis and troubleshooting, in one specific embodiment, the safety interlocking device 2 is also configured to synchronously record image data and system status information at the moment of interlocking.

[0076] This data can be used to assess the accuracy of interlocking, analyze the causes of accidents, or optimize system performance.

[0077] Specifically, the core image data is used for tracing the locking action and verifying the judgment criteria, including: 1. Key frame image for triggering interlocking: including the original captured image and the algorithm-processed image with the composite insulator position and laser hazard area marked, which is the core basis for interlocking triggering; 2. Continuous image sequence before and after locking: reconstructing the complete process from the occurrence of risk to the execution of locking; 3. The most recent closed-loop self-test baseline image before locking is used to verify that the protection system itself is functioning normally before locking.

[0078] System status information is used for response performance verification, compliance review, and troubleshooting, including: 1. Lockout end-to-end timing data: Millisecond-level timestamps of algorithm risk identification, locking command issuance, relay action, and laser power supply cutoff to verify whether the system response speed meets the standard; 2. Lockout core judgment and execution results: composite insulator identification parameters, relay on / off status, laser power-off feedback, and lockout holding status; 3. Operating status of each unit of the device: working status and core parameters of the recognition algorithm, visual acquisition, closed-loop self-test, and interlocking execution module; 4. Status of associated devices: Real-time operating parameters, pose data, and alarm information of lasers, drones / gimbals; 5. Basic operational information: traceability information such as operational routes, personnel, and environment.

[0079] To further enhance the continuous reliability of the system, in one specific embodiment, the safety interlocking device 2 is also configured to automatically trigger a new round of self-inspection process after each continuous operation reaches a preset duration.

[0080] When the UAV laser obstacle removal device 1 operates continuously for extended periods, the system's internal electronic components, sensors, or software may experience performance drift or potential malfunctions due to environmental factors (such as temperature and vibration) or prolonged operation. By setting a preset duration and automatically triggering a new round of self-checks after that duration, the system can ensure that its safety functions remain in optimal working condition throughout the entire operation. This mechanism ensures the system's continuous healthy operation during long-term operations, effectively avoiding performance drift or potential malfunctions that may occur due to prolonged operation, thus providing a more comprehensive and reliable safety guarantee for UAV laser obstacle removal operations.

[0081] Furthermore, the safety interlocking device 2 can integrate a timer module, which starts timing when the operation begins and sends a trigger signal to the closed-loop self-test unit 232 after the preset time is reached, thereby starting the self-test mode.

[0082] Furthermore, the preset duration can be set to 10 minutes.

[0083] This time interval avoids both excessively frequent disruptions and ensures timely system health checks, thereby providing a high level of continuous safety in actual operations. The 10-minute interval is an empirical value designed to balance self-check frequency with operational efficiency, ensuring continuous and effective verification of the safety interlock function during extended operation of the UAV laser obstacle clearer.

[0084] Through the above technical solution, the present invention significantly improves the reliability, response speed and continuous safety of the safety locking device 2 of the UAV laser obstacle clearing device 1.

[0085] In practical applications, the accuracy and robustness of the identification unit 22 in recognizing composite insulators in the image are crucial to ensuring system safety. In complex outdoor operating environments, factors such as changing lighting, diverse viewing angles, partial obstruction, and background interference can challenge the accuracy of image recognition, thereby affecting the timeliness and reliability of the locking action and posing a risk of misidentification or missed identification.

[0086] To address the aforementioned issues, the present invention further optimizes the recognition capability of the recognition unit 22.

[0087] Specifically, the recognition unit 22 incorporates a dedicated image recognition algorithm for the shape, texture, and color features of composite insulators. This dedicated image recognition algorithm employs convolutional neural network image recognition technology. By training the convolutional neural network on composite insulator samples, target detection and localization of composite insulators in the images acquired by the acquisition unit 21 are achieved.

[0088] The dedicated image recognition algorithm built into the recognition unit 22 is used to detect and locate composite insulators in the images acquired by the acquisition unit 21. It can not only identify the presence of composite insulators in the image and determine their category, but also accurately output their spatial position and coordinate information in the image through bounding boxes. This provides a core basis for determining whether the composite insulator is in a laser hazard zone, ensuring the accuracy and timeliness of the locking action.

[0089] Unlike general image recognition solutions, this algorithm is designed and modeled based on the unique visual characteristics of composite insulators, such as their unique skirt structure, silicone rubber surface texture, and typical colors like gray and red, effectively distinguishing composite insulators from background interference and significantly improving recognition accuracy and anti-interference capabilities. Dedicated image recognition algorithms can be implemented using traditional machine learning or deep learning models.

[0090] Furthermore, in this scheme, the dedicated image recognition algorithm specifically adopts Convolutional Neural Network (CNN) image recognition technology. This technology, through multiple convolutional layers, pooling layers, and fully connected layers, can automatically extract discriminative high-level features from the original image without the need for complex manual feature design. It has strong robustness and excellent generalization ability in target detection and image classification tasks, and can effectively cope with complex image changes and improve recognition accuracy.

[0091] Convolutional neural networks (CNNs) are trained using composite insulator samples. Training a CNN is a data-driven, iterative optimization process requiring a large amount of labeled data. Specifically, the training process first involves collecting a massive number of composite insulator images (as positive samples) and images containing either no composite insulators or other interfering objects (as negative samples). These images then undergo precise manual annotation, for example, by drawing bounding boxes to indicate the precise location and category of the composite insulators. Next, these labeled samples are fed into the CNN, where backpropagation algorithms and optimizers (such as Adam and SGD) continuously adjust the network's weights and bias parameters to minimize the error between the model's predictions and the actual annotations. After sufficient training and validation, the network learns and internalizes the inherent visual patterns of composite insulators, enabling it to accurately identify composite insulators even when faced with new and unseen images.

[0092] Therefore, this embodiment of the invention significantly improves the accuracy and robustness of the identification unit 22 in detecting and locating composite insulators in the images acquired by the acquisition unit 21. This combination of specialization and deep learning technology enables the identification unit to effectively cope with complex and ever-changing outdoor operating environments, such as changes in lighting, viewing angle differences, partial occlusion, and background interference, thereby significantly reducing the false recognition rate. When the acquisition unit 21 captures an image of the laser-pointing area, the algorithm can quickly and accurately identify and locate the composite insulators in the image. This ensures that when a composite insulator enters the laser hazard zone, the identification unit 22 can issue a timely and accurate locking command, driving the execution unit 24 to cut off the main power supply circuit of the laser 11, thereby achieving millisecond-level active locking. This effectively prevents the UAV laser obstacle clearing device 1 from accidentally damaging the composite insulators, greatly improving the safety and reliability of UAV laser obstacle clearing operations.

[0093] Please see Figure 1 and Figure 2 In one embodiment, the acquisition unit 21, the identification unit 22, the execution unit 24, and the closed-loop self-test unit 232 are integrated into the same housing to form an intelligent locking module 20.

[0094] By encapsulating the core functional units within a single housing, external connection cables can be effectively reduced, electromagnetic interference can be decreased, and the installation process on the drone can be simplified. The intelligent interlocking module 20, as an independent and fully functional unit, enhances the integration and reliability of the safety interlocking device 2, facilitating production, testing, and maintenance.

[0095] In one embodiment, the diffractive reference plate 231 is composed of a diffractive projection plate and an LED display module (not shown). The LED display module is electrically connected to the closed-loop self-test unit 232, and the diffractive projection plate forms a composite insulator standard feature map.

[0096] The LED display module is driven by a closed-loop self-test unit 232, which provides a controllable light source and enables switching at a predetermined frequency. A standard feature map of a composite insulator is pre-etched or printed on the diffraction projection plate. When the LED display module emits light, this feature map is clearly projected for acquisition by the acquisition unit 21, thus providing a standard second-state self-test image input to the identification unit 22. This design ensures the stability and consistency of the second-state self-test image.

[0097] In one embodiment, the acquisition unit 21 is rigidly connected to the laser 11 to maintain the parallelism between the optical central axis of the acquisition unit 21 and the laser emission path of the laser 11.

[0098] The acquisition unit 21 and the laser 11 are connected by a robust, immovable mechanical connection, such as through an integrated bracket, precision-machined connectors, or shared structural components, ensuring that their relative position remains unchanged under complex conditions such as drone flight, vibration, or external impact. Maintaining the parallelism between the optical central axis and the laser emission path is crucial, as this directly affects the consistency between the visual acquisition area and the laser pointing area, thus ensuring that the identification unit 22 can accurately identify targets that the laser may accidentally damage, such as composite insulators. Therefore, this design ensures a high degree of consistency between the visual acquisition area and the laser pointing area, further improving the accuracy of identification and judgment, thereby comprehensively enhancing the safety performance and operational reliability of the drone laser obstacle clearing device 1.

[0099] Please see Figure 3 The present invention further proposes a safety interlocking method, which is based on the above-mentioned safety interlocking device 2 and specifically includes the following steps: S100. The safety interlocking device 2 is powered on and starts the self-test mode. The closed-loop self-test unit 232 drives the diffractive reference plate 231 to periodically switch between the first state and the second state at a predetermined frequency based on the reference signal to generate a self-test image. The acquisition unit 21 acquires the self-test image synchronously. S200, the recognition unit 22 judges the self-test image, and performs logical self-test verification based on the switching state of the diffractive reference plate 231 and the image recognition result of the self-test image. If the self-test fails, the execution unit 24 is driven to disconnect the main power supply circuit of the laser 11 and report the fault. If the self-test is successful, the monitoring mode in step S300 is entered. S300, Entering real-time monitoring mode, the closed-loop self-testing unit 232 drives the diffractive reference plate 231 to display in the first state based on the self-test success signal, and the acquisition unit 21 continuously acquires on-site images of the laser pointing area in the real operation scenario. S400, the identification unit 22 performs real-time analysis on each frame of the scene image to determine whether there is a composite insulator in the laser hazard area of ​​the scene image; S500: If the composite insulator is detected, a locking command is issued to drive the execution unit 24 to perform a locking action, cutting off the main power supply circuit of the laser 11 to achieve active locking; if the composite insulator is not detected, the laser 11 is kept in an enabled state. S600. During work breaks or when a manual instruction is received, a new round of self-check process is automatically or manually triggered, and steps S100 to S200 are repeated.

[0100] In the aforementioned safety interlocking method, step S100 describes the initial state of the safety interlocking device 2 after power-on and the activation of the self-test mode. When the safety interlocking device 2 is powered on and initialized, the system automatically enters the self-test mode, which is a crucial step to ensure that the entire safety system is in a known good state before any operation begins. In this mode, the closed-loop self-test unit 232 drives the diffractive reference plate 231 to periodically alternate between two states at a specific predetermined frequency according to a preset reference signal, such as an internal clock signal or an external trigger signal. The first state is usually a fully transparent state, simulating the situation where there are no obstacles in the laser pointing area; the second state is the state of projecting a standard feature map of composite insulators, simulating the situation where there are composite insulators in the laser pointing area. The acquisition unit 21 synchronizes with the switching of the diffractive reference plate 231, acquiring the self-test image displayed by the diffractive reference plate 231 in real time. This synchronous acquisition mechanism ensures that the subsequent identification unit can accurately acquire the image data corresponding to the state of the diffractive reference plate 231, providing reliable input for self-test verification.

[0101] Step S200 details the logic and result processing of the self-test verification. After receiving the self-test image acquired by the acquisition unit 21, the identification unit 22 performs image recognition on the self-test image in conjunction with the current switching state of the diffractive reference plate 231. For example, when the diffractive reference plate 231 is in the second state of projecting the standard feature map of the composite insulator, the identification unit 22 should be able to identify the composite insulator features; while when the diffractive reference plate 231 is in the first state of full transparency, it should not identify the composite insulator features. Through this logical matching relationship, the system can perform end-to-end verification of the overall function of the acquisition unit 21 and the identification unit 22. If the self-test verification result does not meet expectations, the self-test is judged to have failed, indicating that there is a fault in the visual recognition or logical judgment link of the safety interlocking device 2. At this time, the identification unit 22 will immediately drive the execution unit 24 to disconnect the main power supply circuit of the laser 11, thereby preventing accidental injury that may be caused by working under fault conditions, and simultaneously report the fault information so that maintenance personnel can handle it in a timely manner. If the self-test is successful, it indicates that the core safety function of the safety interlocking device 2 is normal, and the system will safely enter the monitoring mode described in step S300, ready to carry out the actual obstacle clearing operation.

[0102] Steps S300 and S400 describe the continuous operation process after entering the real-time monitoring mode. Once the self-test is successful, the system enters the real-time monitoring mode. In this mode, the closed-loop self-test unit 232 drives the diffractive reference plate 231 to display in the first state based on the self-test success signal. In the transparent state, it does not interfere with the normal field of vision, ensuring the smoothness of the obstacle clearing operation. The acquisition unit 21 continuously acquires on-site images of the laser-pointing area in the actual operation scenario. Since the optical central axis of the acquisition unit 21 is parallel to the laser emission path of the laser 11, the acquired images can accurately reflect the area where the laser is about to act. The identification unit 22 performs high-speed analysis on each frame of the real-time acquired on-site image to determine whether there are composite insulators in the laser hazard area in the image. The laser hazard area refers to the image area that overlaps with or is adjacent to the laser emission path, which is the range where the laser may act.

[0103] Step S500 specifies the interlocking decision and execution in monitoring mode. If the identification unit 22 identifies a composite insulator within the laser hazard area while analyzing the field image in real time, the identification unit 22 will immediately issue an interlocking command. This command will drive the execution unit 24 to quickly execute the interlocking action, cutting off the main power supply circuit of the laser 11, thereby actively preventing the laser 11 from emitting laser within milliseconds, effectively avoiding accidental damage to the composite insulator. Conversely, if the identification unit 22 does not identify a composite insulator within the laser hazard area, it will keep the laser 11 enabled, allowing the laser 11 to work normally and perform obstacle clearing operations.

[0104] Step S600 further improves the reliability of the safety interlocking method. To ensure the continued reliability of the safety interlocking device 2 before long-term operation or critical operations, it is stipulated that a new round of self-checking process will be automatically or manually triggered during operation breaks or when the system receives a manual trigger command. This means that the system will repeat steps S100 to S200 to verify its own visual acquisition and recognition functions again. This periodic or on-demand self-checking mechanism can promptly detect and handle potential faults that may occur during operation, thereby continuously ensuring the operational safety of the UAV laser obstacle clearing device 1.

[0105] Through the above technical solution, this invention provides a systematic and procedural method for safe interlocking operation. This method ensures that the key functions of the safety interlocking device 2, such as visual acquisition, image recognition, and interlocking execution, are in normal working order before operation begins by forcibly executing a comprehensive self-test process when the safety interlocking device 2 is powered on, effectively avoiding potential accidental injury risks caused by system failures. After a successful self-test, the system seamlessly switches to real-time monitoring mode, continuously performing high-frequency image analysis on the laser-pointing area. Once a composite insulator is identified, a millisecond-level interlocking action is immediately triggered, providing immediate and reliable protection in actual operation. Furthermore, by automatically or manually triggering a new round of self-tests during operation breaks or upon receiving manual commands, this method further ensures the continuous reliability of the safety interlocking device 2 throughout the entire operation cycle, promptly detecting and handling potential system drift or faults, thereby significantly improving the safety and reliability of UAV laser obstacle clearing operations and effectively solving the problem of ensuring the systematic and continuous effective operation of all functions of the safety interlocking device 2.

[0106] Please see Figure 1 and Figure 2 The present invention proposes a drone laser obstacle removal system 100, including a drone laser obstacle removal device 1 and a safety locking device 2 in any of the above embodiments.

[0107] The UAV laser obstacle clearing device 1 includes a laser 11, which includes a laser body 111 and a laser lens 112. The laser lens 112 is disposed on the laser emission path of the laser body 111. The acquisition unit 21, the identification unit 22, the execution unit 24, and the closed-loop self-test unit 232 in the safety interlocking device 2 are integrated in the same housing to form an intelligent interlocking module 20. The intelligent interlocking module 20 and the diffraction reference plate 231 are both disposed above the laser body 111, and the diffraction reference plate 231 is in close contact with the laser lens 112.

[0108] Specifically, the laser body 111 is the main structural part of the laser 11, containing core components such as laser generation, control, and heat dissipation. The laser body 111 is typically made of lightweight, high-strength materials, such as aluminum alloy or carbon fiber composites, to meet the weight and strength requirements of drones. The laser lens 112 is the end optical element of the laser emission path, used to focus or collimate the laser beam, ensuring effective laser energy output and pointing accuracy. The laser lens 112 can be a single lens, a multi-lens group, or a mirror group; its design must comprehensively consider laser wavelength, power density, and environmental adaptability.

[0109] The intelligent interlocking module 20 integrates the acquisition unit 21, the identification unit 22, the execution unit 24, and the closed-loop self-testing unit 232. The intelligent interlocking module 20, together with the diffraction reference plate 231, is positioned above the laser housing 111 to achieve a high degree of system integration and optimized space utilization. Specifically, the intelligent interlocking module 20 can be fixed to the top of the laser housing 111 through an integrated housing. The diffraction reference plate 231 is mounted in front of the intelligent interlocking module 20, ensuring it is within the field of view of the acquisition unit 21. This compact integration helps reduce the overall system size and weight and simplifies the installation process.

[0110] The diffractive reference plate 231 is closely attached to the laser lens 112. This close attachment helps ensure that the acquisition unit 21 can clearly and stably capture the image displayed by the diffractive reference plate 231 during self-testing, reducing ambient light interference and optical distortion, thereby improving the reliability of self-testing.

[0111] By integrating the acquisition unit 21, identification unit 22, execution unit 24, and closed-loop self-test unit 232 of the aforementioned safety interlocking device 2 into the same housing, forming an intelligent interlocking module 20, and placing it together with the diffraction reference plate 231 above the laser body 111, the system achieves high integration and miniaturization, effectively reducing the overall size and weight of the UAV laser obstacle removal system 100, facilitating UAV mounting and flexible operation. More importantly, the diffraction reference plate 231 is positioned close to the laser lens 112, ensuring that the acquisition unit 21 can acquire self-test images with the best viewing angle and clarity in self-test mode, greatly improving the accuracy and reliability of the self-test verification of the self-test module 23, thereby enabling more precise judgment of whether the safety interlocking device 2 is working properly. When the self-test is successful, the system can ensure the safe operation of the laser 11 in monitoring mode; when the self-test fails, it can promptly detect and report the fault, and immediately drive the execution unit 24 to disconnect the main power supply circuit of the laser 11, effectively preventing the risk of accidental injury due to device failure.

[0112] The drone laser obstacle removal system 100 of the above embodiments includes the corresponding safety locking device 2 in any of the above embodiments, and has the beneficial effects of the corresponding device embodiments, which will not be described again here.

[0113] The above description is merely a preferred embodiment of the present invention and the technical principles employed. The present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the claims.

Claims

1. A safety interlocking device, characterized in that, The safety interlocking device is installed on the UAV laser obstacle removal device and includes a data acquisition unit, an identification unit, a self-test module, and an execution unit. The acquisition unit has its optical central axis parallel to the laser emission path of the laser in the UAV laser obstacle removal device, and is used to acquire images of the laser-pointing area. The identification unit is electrically connected to the acquisition unit and is used to determine whether a composite insulator exists in the image acquired by the acquisition unit. The self-test module includes a diffraction reference plate and a closed-loop self-test unit disposed in front of the acquisition unit. The closed-loop self-test unit is electrically connected to the diffraction reference plate and the identification unit, respectively. The closed-loop self-test unit is used to drive the diffraction reference plate to alternately switch between a first state and a second state at a predetermined frequency to generate a self-test image. The first state is a fully transparent state, and the second state is a state of projecting a standard feature map of a composite insulator. The acquisition unit is also used to acquire the self-test image. The execution unit is connected in series in the main power supply circuit of the laser, and its control terminal is controlled by the output signal of the recognition unit. The recognition unit is configured to, in self-test mode, determine whether the safety interlocking device is normal based on the matching relationship between the switching state of the diffractive reference plate and the image recognition result of the self-test image, and control the on / off state of the execution unit based on the judgment result of the self-test mode.

2. The safety interlocking device according to claim 1, characterized in that, The identification unit is also configured to, in monitoring mode, determine whether the UAV laser obstacle clearing device has accidentally damaged the composite insulator based on the image recognition results of the on-site images collected by the acquisition unit, and control the on / off state of the execution unit based on the judgment results of the monitoring mode. The self-test mode is specifically as follows: The closed-loop self-test unit drives the diffractive reference plate to alternately switch between the first state and the second state at a predetermined frequency based on the reference signal to generate a self-test image. The acquisition unit synchronously acquires the self-test image; The identification unit performs a logical self-test based on the switching state of the diffractive reference plate and the identification result of the self-test image. If the self-test fails, the execution unit is driven to disconnect the main power supply circuit of the laser and report the fault. If the self-test succeeds, the unit enters the monitoring mode. The monitoring mode is specifically as follows: The closed-loop self-testing unit drives the diffractive reference plate to display the first state based on the self-test success signal; The acquisition unit acquires real-time images of the laser-pointing area in the actual work scenario; The identification unit analyzes the scene image. If it identifies a composite insulator within the laser hazard area of ​​the scene image, it drives the execution unit to disconnect the main power supply circuit of the laser to achieve active blocking. If it does not identify a composite insulator within the laser hazard area of ​​the scene image, it keeps the laser enabled.

3. The safety interlocking device according to claim 2, characterized in that, In the self-test mode, the recognition unit performs a logical self-test verification based on the switching state of the diffractive reference plate and the image recognition result of the self-test image, including: If the identification unit identifies composite insulator features in the self-test image when the diffractive reference plate is in the second projection state, and does not identify composite insulator features in the self-test image when the diffractive reference plate is in the first state, then the self-test is determined to be successful. If the identification unit fails to identify composite insulator features in the self-test image when the diffractive reference plate is in the second projection state, and / or identifies composite insulator features in the self-test image when the diffractive reference plate is in the first state, then the self-test is determined to have failed.

4. The safety interlocking device according to claim 2, characterized in that, The self-test mode of the identification unit is activated when the safety interlocking device is powered on and initialized, during work breaks, or when a manual trigger command is received. The execution unit is configured to require manual intervention to reset after being disconnected. Before manual intervention, the execution unit maintains the disconnected state of the laser's main power supply circuit.

5. The safety interlocking device according to claim 4, characterized in that, The closed-loop self-testing unit drives the diffractive reference plate to alternately switch between the first state and the second state at a frequency of not less than 5Hz. After the identification unit detects the presence of a composite insulator within the laser hazard area of ​​the on-site image, it sends a locking command to the execution unit within 10ms. Upon receiving the locking command, the execution unit immediately cuts off the main power supply circuit of the laser, and the safety locking device simultaneously records the image data and system status information at the locking moment. The safety interlocking device is configured to automatically trigger a new round of self-inspection process after each continuous operation reaches a preset time.

6. The safety interlocking device according to claim 1, characterized in that, The recognition unit has a built-in dedicated image recognition algorithm for the shape, texture and color features of composite insulators. The dedicated image recognition algorithm adopts convolutional neural network image recognition technology.

7. The safety interlocking device according to claim 1, characterized in that, The acquisition unit, the identification unit, the execution unit, and the closed-loop self-testing unit are integrated into the same housing to form an intelligent locking module; The diffractive reference plate is composed of a diffractive projection plate and an LED display module. The LED display module is electrically connected to the closed-loop self-test unit. The diffractive projection plate forms a standard feature map of a composite insulator. The execution unit includes a fast relay, the control terminal of which is controlled by the output signal of the identification unit; The acquisition unit is rigidly connected to the laser to maintain the parallelism between the optical central axis of the acquisition unit and the laser emission path of the laser.

8. A safety interlocking method, characterized in that, Based on a safety interlocking device according to any one of claims 1 to 7, the safety interlocking method includes: S100. The safety interlocking device is powered on and starts the self-test mode. The closed-loop self-test unit drives the diffractive reference plate to periodically switch between the first state and the second state at a predetermined frequency based on the reference signal to generate a self-test image. The acquisition unit synchronously acquires the self-test image. S200: The recognition unit judges the self-test image, performs logical self-test verification based on the switching state of the diffractive reference plate and the image recognition result of the self-test image. If the self-test fails, the execution unit is driven to disconnect the main power supply circuit of the laser and report the fault. If the self-test is successful, the monitoring mode in step S300 is entered. S300, Enter real-time monitoring mode. The closed-loop self-testing unit drives the diffractive reference plate to display in the first state based on the self-test success signal. The acquisition unit continuously acquires on-site images of the laser pointing area in the real operation scenario. S400, The identification unit performs real-time analysis on each frame of the scene image to determine whether there is a composite insulator within the laser hazard area of ​​the scene image; S500: If the composite insulator is detected, a locking command is issued to drive the execution unit to perform a locking action, cutting off the main power supply circuit of the laser to achieve active locking; if the composite insulator is not detected, the laser is kept in an enabled state. S600. During work breaks or when a manual instruction is received, a new round of self-check process is automatically or manually triggered, and steps S100 to S200 are repeated.

9. The safety interlocking method according to claim 8, characterized in that, In the self-test mode, the recognition unit performs a logical self-test verification based on the switching state of the diffractive reference plate and the image recognition result of the self-test image, including: If the identification unit identifies composite insulator features in the self-test image when the diffractive reference plate is in the second projection state, and does not identify composite insulator features in the self-test image when the diffractive reference plate is in the first state, then the self-test is determined to be successful. If the identification unit fails to identify composite insulator features in the self-test image when the diffractive reference plate is in the second projection state, and / or identifies composite insulator features in the self-test image when the diffractive reference plate is in the first state, then the self-test is determined to have failed.

10. A drone laser obstacle removal system, characterized in that, include: A drone laser obstacle clearing device and a safety locking device as described in any one of claims 1 to 7; The UAV laser obstacle removal device includes a laser, which includes a laser body and a laser lens. The laser lens is disposed in the laser emission path of the laser body. The acquisition unit, identification unit, execution unit, and closed-loop self-test unit in the safety interlocking device are integrated into the same housing to form an intelligent interlocking module. The intelligent interlocking module and the diffraction reference plate are both located above the laser body, and the diffraction reference plate is in close contact with the laser lens.

Images