Inspection robot, inspection method and related system for photovoltaic backsheet

Abstract

The application provides an inspection robot, an inspection method and a related system for a photovoltaic backboard. A main processor of the inspection robot is configured to receive and respond to a first walking instruction, send a first posture adjustment instruction to a rotating lifting mechanism, and send a first data acquisition instruction to a detection module. A walking mechanism is configured to receive and respond to the first walking instruction to perform forward and backward movement on a guide piece. The rotating lifting mechanism is configured to receive and respond to the posture adjustment instruction. The detection module is configured to receive and respond to the first data acquisition instruction to acquire data. The main processor is further configured to receive first data, detect whether there is a first abnormal position according to the first data, perform data acquisition and abnormality detection operations on a first end face if it is detected that there is a first abnormal position, and send a posture adjustment instruction to the rotating lifting mechanism if it is detected that there is no first abnormal position. In this way, accurate identification of hidden dangers of the photovoltaic backboard and inspection efficiency are achieved.

Description

Technical Field

[0001] This application relates to the field of robotics technology, and in particular to an inspection robot, inspection method and related system for photovoltaic backsheets. Background Technology

[0002] During use, photovoltaic (PV) backsheets face issues such as poor soldering of PV connector joints, poor soldering of junction boxes, and backsheet aging. These problems can lead to safety and property damage during operation. For example, poor soldering of PV connector joints can easily cause grounding discharge and arcing, which can even cause fires in severe cases. Regular inspections of operating PV backsheets can reduce the occurrence of such problems. Currently, PV backsheet inspections are typically conducted manually or using handheld portable infrared inspection equipment, and these methods have achieved some success.

[0003] However, due to the narrow space on the corrugated steel roofs where photovoltaic backsheets are installed, the obstruction by brackets, and the inaccessibility of cable joints, manual inspection methods are difficult to reach. In addition, for large-scale corrugated steel roof projects, maintenance personnel cannot quickly locate potential hazards with the naked eye and can only mark them and notify the back office for centralized troubleshooting, which results in low inspection efficiency.

[0004] Therefore, improving inspection efficiency and accuracy in identifying potential problems in photovoltaic backsheets are urgent issues to be addressed during the inspection process. Summary of the Invention

[0005] This application provides an inspection robot, inspection method, and related system for photovoltaic backsheets, which improves the efficiency of inspection and the accuracy of identifying potential problems in photovoltaic backsheets.

[0006] In a first aspect, embodiments of this application provide an inspection robot for photovoltaic backsheets. The inspection robot includes: a main processor, a walking mechanism, a rotating and lifting mechanism, a detection module mounted on the rotating and lifting mechanism, and a guide component mounted on a first end face of the photovoltaic backsheet. The guide component is at least one of a guide rail or a steel wire rope. The first end face is the opposite side of the photovoltaic backsheet facing the light-receiving surface. The main processor is used to receive and respond to the first walking command for the inspection robot, send the first posture adjustment command to the rotating lifting mechanism, and send the first data acquisition command to the detection module. The walking mechanism is used to receive the first walking command and, in response to the first walking command, to move forward and / or backward on the guide member. The rotary lifting mechanism is used to receive and respond to the attitude adjustment command; The detection module is used to receive and respond to the first data acquisition command; acquire first data from the first end face; and send the first data to the main processor. The main processor is used to receive the first data and detect whether there is a first abnormal position on the first end face based on the first data; If the first abnormal location is detected, wait for a preset time and then continue to perform data acquisition and abnormal detection operations on the first end face; and / or send abnormal information generated in response to the abnormal detection operation to the cloud platform. If the first abnormal position is not detected, the attitude adjustment command is sent to the rotating lifting mechanism, and / or the second walking command is sent to the walking mechanism.

[0007] Secondly, embodiments of this application provide an inspection method for photovoltaic backsheets, applied to the main processor of an inspection robot. The inspection robot includes: a walking mechanism, a rotating and lifting mechanism, a detection module mounted on the rotating and lifting mechanism, and a guide component mounted on a first end face of the photovoltaic backsheet. The guide component is at least one of a guide rail or a steel wire rope. The first end face is the opposite side of the photovoltaic backsheet facing the light-receiving surface. The method includes: Receive and respond to the first walking command for the inspection robot; Send a first attitude adjustment command to the rotary lifting mechanism; and send a first data acquisition command to the detection module; The detection module is controlled to receive and respond to the first data acquisition command; Collect first data from the first end face; and send the first data to the main processor; Receive the first data, and detect whether there is a first abnormal position on the first end face based on the first data; If the first abnormal location is detected, wait for a preset time and then continue to perform data acquisition and abnormal detection operations on the first end face; and / or send abnormal information generated in response to the abnormal detection operation to the cloud platform. If the first abnormal position is not detected, the attitude adjustment command is sent to the rotating lifting mechanism, and / or the second walking command is sent to the walking mechanism.

[0008] Thirdly, embodiments of this application provide an inspection system for photovoltaic backsheets, which can be used to execute instructions for the steps in any of the methods in the second aspect of embodiments of this application.

[0009] Fourthly, embodiments of this application provide a robot, including a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs include instructions for performing steps in any method of the second aspect of this application.

[0010] Fifthly, embodiments of this application provide a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in any method of the first aspect of this application. The computer program product may be a software installation package.

[0011] By implementing the embodiments of this application, the following beneficial effects are achieved: This application provides an inspection robot for photovoltaic backsheets. The inspection robot includes: a main processor, a walking mechanism, a rotating and lifting mechanism, a detection module mounted on the rotating and lifting mechanism, and a guide component mounted on a first end face of the photovoltaic backsheet. The first end face is the opposite side of the photovoltaic backsheet facing the light source. The main processor receives and responds to a first walking command for the inspection robot, sends a first posture adjustment command to the rotating and lifting mechanism, and sends a first data acquisition command to the detection module. The walking mechanism receives the first walking command and responds by moving forward and / or backward on the guide component. The rotating and lifting mechanism... The system includes a receiving and responding attitude adjustment command; a detection module for receiving and responding to a first data acquisition command; acquiring first data from the first end face; and sending the first data to the main processor; the main processor is also used to receive the first data, detect whether there is a first abnormal position on the first end face based on the first data, and if a first abnormal position is detected, wait for a preset time and then continue to perform data acquisition and abnormal detection operations on the first end face; and / or send abnormal information generated in response to the abnormal detection operation to the cloud platform, and if no first abnormal position is detected, send an attitude adjustment command to the rotating lifting mechanism, and / or send a second walking command to the walking mechanism. Thus, by arranging steel wire ropes or guide rails (guide components) on the photovoltaic backsheet and mounting an inspection robot equipped with visible light cameras, infrared cameras, and a multi-degree-of-freedom gimbal, image acquisition, temperature monitoring, and intelligent recognition of the photovoltaic connector joints, junction boxes, and photovoltaic backsheet status can be achieved, enabling accurate identification of potential problems on the photovoltaic backsheet and thus improving the efficiency of photovoltaic backsheet inspection. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments 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.

[0013] Figure 1 This is an architecture diagram of an inspection system for photovoltaic backsheets provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a robot provided in an embodiment of this application; Figure 3 This is a flowchart illustrating an inspection method for photovoltaic backsheets provided in an embodiment of this application. Figure 4 This is a front view of an inspection robot inspecting a photovoltaic backsheet, as provided in an embodiment of this application. Figure 5 This is a schematic diagram of the structure of a walking module provided in an embodiment of this application; Figure 6 This is a schematic diagram of a scene where an inspection robot inspects a photovoltaic backsheet, as provided in an embodiment of this application. Figure 7 This is a schematic diagram of the installation structure of an inspection robot on a photovoltaic backsheet according to an embodiment of this application; Figure 8 This is a block diagram of an inspection robot for photovoltaic backsheets provided in an embodiment of this application. Detailed Implementation

[0014] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0015] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0016] It should be understood that the term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this document indicates that the preceding and following related objects are in an "or" relationship. In the embodiments of this application, "multiple" refers to two or more.

[0017] In the embodiments of this application, "at least one item" or its similar expression refers to any combination of these items, including any combination of a single item or a plurality of items. "One or more" means one or more, while "multiple" means two or more. For example, "at least one item" of a, b, or c can represent the following seven cases: a, b, c; a and b; a and c; b and c; a, b, and c. Each of a, b, and c can be an element or a set containing one or more elements.

[0018] In this application, the term "connection" refers to various connection methods, such as direct connection or indirect connection, to achieve communication between devices. This application does not impose any limitations on this.

[0019] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0020] The following is an explanation of the relevant terms used in this application: Photovoltaic connector: Commonly known as MC4 connector in the photovoltaic field, it is typically a multi-contact connector. The MC4 connector is a type of connector used in solar photovoltaic systems, usually for connecting cables between solar panels and solar inverters or other electrical equipment.

[0021] Solar panel backsheets are prone to problems such as faulty MC4 connector soldering and material aging during operation, which can easily lead to fires and property damage. Regular inspections of solar panel backsheets can reduce the occurrence of fires and other such incidents. Current inspection methods involve maintenance personnel relying on experience or manually using handheld infrared inspection equipment. However, due to the narrow space of corrugated steel roofs, obstructions from supports, and inaccessible cable connectors, coupled with the difficulty for maintenance personnel to pinpoint potential hazards based on experience, blind spots exist in the inspections, potentially causing some hazards to go undetected and affecting the accuracy of hazard detection. Furthermore, for large-scale corrugated steel roof projects, manual or handheld inspection methods are inefficient.

[0022] To address the aforementioned problems, this application provides an inspection robot, inspection method, and related system for photovoltaic backsheets. The method is applied to the main processor of the inspection robot, which includes: a walking mechanism, a rotating and lifting mechanism, a detection module mounted on the rotating and lifting mechanism, and a guide component mounted on the first end face of the photovoltaic backsheet. The guide component is at least one of a guide rail or a steel wire rope. The first end face is the opposite side of the photovoltaic backsheet facing the light source. The method includes: receiving and responding to a first walking command for the inspection robot; sending a first posture adjustment command to the rotating and lifting mechanism; and sending a first data acquisition command to the detection module to control the detection module. The system receives and responds to the first data acquisition command, acquires first data from the first end face, sends the first data to the main processor, receives the first data, detects whether there is a first abnormal position on the first end face based on the first data, and if the first abnormal position is detected, waits for a preset time before continuing to perform data acquisition and abnormal detection operations on the first end face; and / or sends abnormal information generated in response to the abnormal detection operation to the cloud platform; if the first abnormal position is not detected, sends the attitude adjustment command to the rotating lifting mechanism, and / or sends a second walking command to the walking mechanism. In this way, the efficiency of the photovoltaic backsheet inspection is improved, as is the accuracy of identifying potential problems with the photovoltaic backsheet.

[0023] The following is combined Figure 1 The system architecture of an inspection method for photovoltaic backsheets according to an embodiment of this application is described. Figure 1 This is an architecture diagram of an inspection system for photovoltaic backsheets provided in an embodiment of this application. The inspection system 100 for photovoltaic backsheets includes a server 110 and a photovoltaic inspection robot 120.

[0024] The server 110 is connected to the photovoltaic inspection robot 120. The server 110 is used to send inspection task instructions, attitude adjustment instructions, etc. to the photovoltaic inspection robot 120, and to receive abnormal information or inspection data uploaded by the photovoltaic inspection robot 120. The server 110 can also send abnormal information to the operation and management personnel.

[0025] The photovoltaic inspection robot 120 is positioned on the non-sunlit side of the photovoltaic backsheet and performs automatic inspections along a preset guide path according to tasks issued by the server 110. The photovoltaic inspection robot 120 includes a main processor 121, a walking mechanism 122, a rotating and lifting mechanism 123, and a detection module 124. The main processor 121 is used to parse the motion control information issued by the server 110 and generate local control commands for adjusting the robot's posture, triggering data acquisition, and uploading data, thereby achieving real-time management and scheduling of inspection tasks. In one possible embodiment, the main processor 121 has multi-threaded task scheduling and edge intelligent computing capabilities to ensure stable operation in narrow rooftop environments. The walking mechanism 122 is installed at the bottom of the photovoltaic inspection robot 120 and is used for forward and backward movement on guide rails or steel cables set on the photovoltaic backsheet, thereby supporting the robot to achieve full-coverage inspection throughout the entire photovoltaic array. The walking mechanism 122 can improve its anti-tilt capability by forming a clamping structure with the guide components, ensuring that the robot can maintain stable movement under wind pressure disturbances. The rotating and lifting mechanism 123 is used to execute rotational and lifting actions according to the attitude adjustment instructions issued by the main processor 121, so that the detection module 124 can perform detailed observation of key parts of the photovoltaic backsheet from different spatial angles. The detection module 124 includes a visible light camera, an infrared camera, and a communication module, used to acquire visible light images and thermal imaging data of the backsheet surface according to the trigger instructions of the main processor 121. In one possible embodiment, the infrared camera of the detection module 124 can achieve a temperature measurement accuracy of ±0.1℃, thereby accurately capturing typical fault characteristics such as poor soldering of photovoltaic connectors and localized overheating of junction boxes.

[0026] As can be seen, through the collaborative operation of server 110 and photovoltaic inspection robot 120, the photovoltaic backsheet inspection system 100 can automatically detect suspected hidden dangers on photovoltaic backsheets, upload data, and perform remote analysis. This avoids the limitations of traditional manual inspections caused by narrow spaces and bracket obstructions, improving the accuracy of hidden danger identification and inspection efficiency. Furthermore, this architecture enables task distribution and data feedback through a communication network. Combined with the robot's multimodal detection capabilities, it can achieve long-term stable operation in the complex environment of photovoltaic power plants.

[0027] The following is combined Figure 2 The robot in the embodiments of this application will be described. Figure 2 This application provides a schematic diagram of the structure of a robot, as shown in the embodiment of the present application. Figure 2 As shown, the robot 200 includes one or more processors 210, a memory 220, a communication interface 230, and one or more programs 221. The processor 210 is connected to the memory 220 and the communication interface 230 via an internal communication bus.

[0028] The one or more programs 221 are stored in the memory 220 and configured to be executed by the processor 210. The one or more programs 221 include instructions for performing any step in the above method embodiments.

[0029] The processor 210 can be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, units, and circuits described in conjunction with the disclosure of this application. The processor can also be a combination that implements computational functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc. The communication unit can be a communication interface, transceiver, transceiver circuit, etc., and the storage unit can be a memory.

[0030] The memory 220 can be volatile memory or non-volatile memory, or it can include both. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0031] It is understood that robot 200 may include more or fewer structural elements than those shown in the above structural block diagram, such as a power module, physical buttons, a Wi-Fi module, a speaker, a Bluetooth module, sensors, a display module, etc., without limitation. It is understood that robot 200 may be equipped with... Figure 1 The architecture of the inspection system for photovoltaic backsheets.

[0032] After understanding the software and hardware architecture of this application, the following will be combined with... Figure 3 This application describes an inspection method for photovoltaic backsheets. Figure 3 This is a flowchart illustrating an inspection method for photovoltaic backsheets provided in an embodiment of this application. The method is applied to the main processor of an inspection robot, which includes: a walking mechanism, a rotating and lifting mechanism, a detection module mounted on the rotating and lifting mechanism, and a guide component mounted on the first end face of the photovoltaic backsheet. The guide component is at least one of a guide rail or a steel wire rope. The first end face is the opposite side of the photovoltaic backsheet facing the light-receiving surface. The method specifically includes the following steps: Step S310: Receive and respond to the first walking command for the inspection robot.

[0033] For a better understanding of the description of the following embodiments, please refer to Figure 4 , Figure 4 This is a front view of an inspection robot inspecting a photovoltaic backsheet, as provided in an embodiment of this application. As can be seen in the front view 400, the inspection robot mainly consists of a walking mechanism, a rotating and lifting mechanism, and a detection module. These components work together to achieve automatic inspection and anomaly detection of the photovoltaic backsheet. The inspection robot is positioned below the photovoltaic backsheet and moves directionally along the length of the backsheet using guides. These guides can be rails or steel cables; preferably, they are a pair of parallel steel cables. Specifically, the photovoltaic backsheet is typically installed on a corrugated steel roof or a support structure, and photovoltaic connectors, junction boxes, and other electrical components are fixed on it. The inspection robot is positioned below the photovoltaic backsheet, maintaining a safe clearance to avoid mechanical contact or obstruction of the photovoltaic modules. The steel cables are arranged along the underside of the photovoltaic backsheet to provide a stable movement path for the inspection robot, enabling precise positioning along the inspection path. The guides can be rails or steel cables to adapt to different installation conditions in various photovoltaic scenarios. The walking mechanism is located at both ends of the inspection robot, and its main function is to support the robot's forward and backward movement on the guide. The walking mechanism includes a set of wheels, a drive unit, and a clamping structure to ensure stable operation of the robot within the narrow space under the back panel, unaffected by wind or tilting. A rotating and lifting mechanism is located in the central area of ​​the inspection robot. This mechanism drives the detection module to adjust its posture in the vertical and horizontal directions. Through rotation and lifting, the rotating and lifting mechanism allows the detection module to observe different parts of the photovoltaic back panel at fixed angles and distances, which is particularly suitable for image acquisition and temperature monitoring of critical components such as junction boxes and photovoltaic connectors. This structural design maintains a stable field of view even in complex environments such as bracket obstruction and changes in back panel height. The detection module is installed at the end of the rotating and lifting mechanism and is the unit for the inspection robot to perform data acquisition tasks. The detection module includes an infrared camera and a visible light camera to acquire thermal imaging data and visible light images of the photovoltaic back panel, respectively, for subsequent temperature analysis, appearance identification, and hazard assessment. Meanwhile, the detection module can be further integrated with a communication module to send the collected data to the main processor or server for processing in real time.

[0034] The first movement command controls the inspection robot to move forward or backward on the wire rope or guide rail. The inspection robot communicates with the server (cloud platform). When the inspection robot is inspecting the back of the photovoltaic backsheet, the server sends a movement command, i.e., the first movement command, to the inspection robot. After receiving the first movement command, the main processor of the inspection robot controls the movement mechanism of the inspection robot to move on the wire rope or guide rail.

[0035] Specifically, after receiving the first walking command, the main processor first calibrates the current posture, position, and clamping state of the walking mechanism to ensure a reliable guiding relationship between the walking wheels and the guide. When the guide is a steel cable, the main processor can also access sensor data from the tension monitoring component to confirm that the tension of the steel cable is within a safe range capable of supporting the weight of the inspection robot. Next, the main processor sends the first walking command to the drive module of the walking mechanism. The drive module adjusts the rotational speed, direction, and start / stop method of the walking wheels according to this command, thereby driving the inspection robot to move along the guide towards the predetermined inspection starting position. During the walking process, the inspection robot uses the posture detection module to provide real-time feedback on whether it deviates from the guide or tilts, and the main processor performs dynamic corrections to ensure the stability and continuity of the walking motion.

[0036] It should be noted that the first walking command not only initiates the inspection robot's initial movement but also serves as a template for its basic control commands. Throughout multiple inspection cycles, the main processor can adjust the first walking command based on real-time data, anomaly records, and inspection coverage statistics to create an optimized inspection path. Furthermore, in practical use, the inspection robot may need to reduce its walking speed before entering critical inspection locations to allow for subsequent attitude adjustments and high-precision data acquisition. Therefore, the first walking command can include parameters such as walking speed, starting acceleration, and deceleration threshold, and can be adaptively configured according to the specific structural characteristics of the photovoltaic power station.

[0037] Step S320: Send a first attitude adjustment command to the rotary lifting mechanism; and send a first data acquisition command to the detection module.

[0038] The first attitude adjustment command is generated by the main processor based on the structural characteristics of the photovoltaic modules, the potential distribution patterns of anomalies, and the actual installation angle of the guide components in the inspection task scenario. Since photovoltaic modules on corrugated steel roofs are usually installed at a fixed tilt angle, and the non-sunlit side of the photovoltaic backsheet has various obstruction structures such as brackets, cables, and junction boxes, the main processor needs to evaluate the optical axis angle, vertical height, and distance from the backsheet surface of the current detection module through the attitude model before the inspection robot is about to enter the designated inspection position, so as to determine whether rotation adjustment, lifting adjustment, or a combination of both is required.

[0039] Specifically, when the inspection robot responds to the first walking command and moves to the vicinity of the target inspection position, the main processor first determines whether there is a deviation in the shooting angle of the detection module based on the feedback data from the robot's attitude sensor (which can be a gyroscope, angle sensor, or lifting position encoder, which is not limited here). For example, when the robot is below the height of the MC4 connector or junction box, the main processor will generate a lifting command to drive the lifting module to raise the detection module to a suitable height to ensure that the visible light camera and infrared camera can completely cover the outer shell area of ​​the connector; when the robot is at an angular deviation from the back panel, the main processor will generate a rotation command to adjust the horizontal angle of the detection module so that the camera optical axis is approximately perpendicular to the back panel surface to reduce image distortion and thermal imaging bias. After sending the attitude adjustment command, the main processor sends a first data acquisition command to the detection module. This command includes two types of information: the first is an image acquisition trigger signal for the visible light camera; the second is a thermal imaging acquisition command for the infrared camera. After receiving the acquisition command, the control unit inside the detection module will drive the visible light imaging unit and the infrared detection unit to perform synchronous acquisition respectively. The synchronous acquisition method can ensure that the image data and thermal imaging data have the same spatial reference position, thereby avoiding errors in subsequent image processing and position mapping.

[0040] It is evident that, on the one hand, by performing posture adjustment before the robot enters the inspection area, the inspection module can obtain clear, stable, and unobstructed images and thermal data, avoiding blind spots caused by limited angles and severe obstruction in traditional manual inspection. On the other hand, by collecting data after posture adjustment, the thermal imaging and visible light images maintain spatiotemporal consistency, providing a reliable foundation for subsequent temperature anomaly judgment and visual feature comparison. Thus, the inspection robot can not only adapt to the high-density, complex support structure of photovoltaic backsheets, but also achieve accurate data collection for different fault modes such as MC4 connectors, junction boxes, and backsheet material aging.

[0041] Step S330: Control the detection module to receive and respond to the first data acquisition command.

[0042] The detection module includes a visible light camera, an infrared camera, and a communication module connected to them. Its main function is to synchronously sense the temperature and morphology of the non-illuminated side of the photovoltaic backsheet. After the main processor completes attitude adjustment and sends the first data acquisition command, it needs to further control the detection module to correctly execute the command and activate the corresponding imaging unit internally. This ensures that the acquired image and thermal imaging data are consistent in the time dimension, thus providing a reliable foundation for subsequent processing such as temperature anomaly screening, image segmentation, and feature parameter extraction.

[0043] Specifically, after receiving the first data acquisition command from the main processor, the detection module activates the visible light camera and sets the exposure time, frame rate, focal length, and aperture parameters according to the command, enabling it to obtain a sufficiently clear image under low-brightness conditions on the non-illuminated side of the photovoltaic backsheet. Simultaneously, the infrared camera calibrates its internal temperature reference curve according to the thermal imaging acquisition command and activates a microbolometer array or thermistor to capture the temperature distribution at multiple locations on the back of the photovoltaic module. During the acquisition process, the visible light camera and the infrared camera need to maintain synchronized triggering and exposure to ensure that image data and thermal imaging data are acquired at the same location and at the same time.

[0044] Step S340: Collect first data from the first end face; and send the first data to the main processor.

[0045] The first data includes first image data and first thermal imaging data reflecting the operating status of the back side (non-sunlit side) of the photovoltaic backsheet. The data sources correspond to the visible light camera and infrared camera in the detection module, respectively. The purpose of acquiring the first data is to obtain the actual temperature distribution and image details of key components on the back side of the photovoltaic backsheet (including the MC4 connector, junction box, and backsheet substrate area) during inspection, so as to facilitate the main processor to perform subsequent anomaly detection, image segmentation, and status parameter evaluation.

[0046] Specifically, after responding to the first data acquisition command, the detection module first activates the visible light camera to perform a full-field scan of the first end face. During the scan, the visible light camera acquires a high-resolution image of the back panel area based on the optical parameters (including focal length, exposure, and aperture size) previously set by the main processor. This image includes not only the appearance of the MC4 connector and junction box, but also visual features such as discoloration, cracks, bulging, or aging marks that may appear on the back panel surface. Simultaneously, the infrared camera performs thermal imaging acquisition on the same area at the same time. By detecting the infrared radiation energy distribution on the back panel surface, it obtains a temperature matrix of multiple sampling points, forming the first thermal imaging data. To ensure data accuracy, the infrared camera performs temperature drift compensation and background temperature calibration before acquisition, thereby reducing errors caused by ambient temperature. After data acquisition, the detection module preprocesses the image data and thermal imaging data, including formatting, noise reduction, pixel calibration, and timestamp synchronization. Subsequently, the detection module sends the first data to the main processor through its internal communication module.

[0047] It is evident that by collecting multi-source data from key components of the photovoltaic backsheet and transmitting this data to the main processor in real time and reliably, the quality of input data in the anomaly identification stage of the inspection system can be effectively improved. This enables temperature anomaly judgment, image feature matching, and state parameter calculation to be carried out on a complete, synchronous, and reliable data basis, thereby significantly improving the accuracy of the inspection results.

[0048] Step S350: Receive the first data and detect whether there is a first abnormal position on the first end face based on the first data.

[0049] The first data includes first image data and first thermal imaging data, which are synchronously acquired by the detection module and sent to the main processor.

[0050] Specifically, after the main processor receives the first data from the detection module, it first performs temperature matrix analysis on the first thermal imaging data. This analysis includes spatial mapping, region division, and temperature feature extraction for the temperature of each sampling point. The main processor filters out target temperature points with significantly higher temperatures based on a preset temperature threshold (which can be set empirically according to the MC4 or junction box manual and sunlight conditions), and maps these points to the first image data at the same location, thereby obtaining the target image of the abnormal area. Next, the main processor performs image segmentation based on the target image, extracting the local image regions corresponding to the photovoltaic connector, junction box, and backplane, and further analyzes the color, texture, and shape features of the image. Taking the MC4 connector as an example, poor soldering or contact is usually accompanied by local discoloration, blackening, or surface aging of the connector shell; while overheated areas of the junction box may appear as obvious hot spots in infrared images and show shell deformation or edge cracking in visible light images. The main processor determines whether there is a deviation from the normal state at that location by comparing the difference between the parameter features of the target image and the pre-stored normal sample features. When the difference exceeds a set threshold, the location is marked as an anomaly, i.e., the first anomaly location. If the target image does not show any abnormal characteristics after image segmentation and parameter recognition, the main processor determines the location as a normal area and can simultaneously upload the temperature and image features of the location to the cloud platform for long-term evaluation of the operating status by photovoltaic operators.

[0051] As can be seen, by integrating thermal imaging and visible light imagery into a single modal data analysis method, the inspection robot can accurately identify various potential problems such as MC4 solder joint defects, junction box overheating, and backplane deformation without relying on human experience. Temperature screening reduces computational load and improves detection efficiency; visual feature comparison enhances recognition dimensions and improves accuracy. Ultimately, this allows the inspection robot to maintain high reliability and accuracy even in the complex and heavily shaded environment behind photovoltaic modules, providing safer and smarter operation and maintenance support for photovoltaic power plants.

[0052] In one possible embodiment, the first data includes first image data and first environmental data. The step of detecting whether a first abnormal location exists on the first end face based on the first data specifically includes the following steps: 351. Extract the temperatures corresponding to multiple locations in the first end face from the first environmental data to obtain multiple temperatures; each of the multiple temperatures corresponds one-to-one with each of the multiple locations; 352. When the target temperature is greater than a preset temperature threshold, the image of the target location corresponding to the target temperature is selected from the first image data to obtain the target image; the target temperature is any one of the plurality of temperatures; 353. Based on a preset image segmentation algorithm, the target location in the target image is segmented to obtain the image segmentation result; 354. Determine the state parameters of the target location based on the image segmentation results to obtain the target state parameters; 355. If the target state parameter indicates an abnormal state, the target position shall be recorded as the first abnormal position; 356. If the target state parameter indicates a normal state, then the target state parameter is sent to the cloud platform.

[0053] The first environmental data is a temperature data matrix obtained by an infrared camera or thermal imaging unit, which is used to reflect the real-time heat distribution on the non-illuminated side of the photovoltaic backsheet; the first image data is image information collected by a visible light camera, which is used to provide structured visual information on the surface condition of the photovoltaic connector, junction box and backsheet.

[0054] Specifically, the first step is to analyze the initial environmental data. Since thermal imaging data is typically presented as a two-dimensional temperature matrix, with each temperature point in the matrix corresponding one-to-one with its location in the actual physical space, the main processor can extract the temperature values ​​at multiple locations within this matrix based on the coordinate calibration information obtained during acquisition. After extraction, the main processor can construct a temperature set containing multiple temperature points to determine whether different areas of the photovoltaic backsheet have temperatures deviating from the normal operating range. After obtaining multiple temperatures, each temperature point is compared with a preset temperature threshold. When a target temperature exceeds the threshold, it indicates a potential hazard at that location. The main processor then filters out the target image corresponding to the coordinate location from the initial image data. Because image acquisition and thermal imaging acquisition are performed synchronously, they can achieve precise spatial mapping. After obtaining the target image, it is further processed based on a preset image segmentation algorithm. This algorithm can employ edge feature-based methods, region growing, deep learning, or color clustering to accurately extract the connector housing structure, junction box edge contour, and local backplate regions from the overall image. The main purpose of this segmentation operation is to remove complex backgrounds from the image, retaining only key areas for state analysis and improving the accuracy of subsequent feature recognition. After obtaining the image segmentation results, the main processor can extract state parameters from the segmented regions. State parameters can include color feature parameters (such as RGB distribution), texture feature parameters (such as gray-level co-occurrence matrix features), and shape feature parameters (such as outer contour regularity and surface unevenness). By analyzing these parameters, the main processor can determine whether the connector housing has discolored, whether the wiring is loose, whether the junction box housing is damaged, and whether the backplate shows signs of deformation, bulging, or aging. If the target state parameters are abnormal, such as a significantly darker color, deformed outer contour of the housing, disordered texture, or typical thermal damage characteristics, the target location is marked as the first abnormal location, triggering further verification or alarm procedures. If the target status parameter is normal, meaning that there are no suspicious deviations in the image and temperature, the parameter can be uploaded to the cloud platform as a normal inspection record to build historical data of the power plant's operating status, so as to carry out subsequent equipment health assessment, long-term trend analysis and risk prediction.

[0055] It is evident that the combined analysis method of ambient temperature screening, image screening, image segmentation, and state parameter recognition has enabled high-precision positioning of potential anomalies in photovoltaic backsheets. Temperature screening ensures inspection efficiency, image segmentation improves the accuracy of anomaly identification, and state parameter judgment enables multi-dimensional intelligent analysis.

[0056] In one possible embodiment, determining the state parameters of the target location based on the image segmentation result to obtain the target state parameters specifically includes the following steps: 3541. Obtain images of a normal photovoltaic connector and a normal junction box; 3542. Extract the image segmentation results for the photovoltaic connector and the junction box from the image segmentation results respectively to obtain the photovoltaic connector image and the junction box image; 3543. Determine the image parameters corresponding to the photovoltaic connector image and the junction box image to obtain the first photovoltaic connector image parameters and the first junction box image parameters; 3544. Determine the image parameters corresponding to the normal photovoltaic connector image and the normal junction box image to obtain the second photovoltaic connector image parameters and the second junction box image parameters; 3545. Determine the state parameters of the photovoltaic connector based on the first photovoltaic connector image parameters and the second photovoltaic connector image parameters to obtain the photovoltaic connector state parameters; 3546. Determine the state parameters of the junction box based on the first junction box image parameters and the second junction box image parameters to obtain the junction box state parameters; 3547. Extract the image data of the back plate in the first end face from the image segmentation result to obtain the back plate image; 3548. Extract deformation features from the backplate image to obtain deformation parameters; 3549. Determine the target state parameter based on the deformation parameter, the junction box state parameter, and the photovoltaic connector state parameter.

[0057] The images of normal photovoltaic connectors and junction boxes refer to images without any abnormal conditions such as loose connections, damage, or aging. These images can be obtained from the cloud platform by the inspection robot. After acquiring the normal images, the independent image regions of the photovoltaic connectors and junction boxes are extracted from the overall image of the photovoltaic backsheet processed by the image segmentation algorithm, using edge detection to reduce interference information from other areas of the photovoltaic backsheet, resulting in target images containing only the photovoltaic connectors and junction boxes. Next, image parameters are extracted, including geometric parameters and color parameters in the visible light dimension and temperature distribution parameters in the infrared dimension. For the extracted photovoltaic connector and junction box images, indicators such as mating gap, shell flatness, RGB color values, and joint temperature rise gradient are obtained, namely the first photovoltaic connector image parameters and the first junction box image parameters. At the same time, quantitative indicators of the same dimension are extracted from the normal sample images to form the second photovoltaic connector image parameters and the second junction box image parameters. After parameter extraction, the two types of parameters are compared with thresholds using difference calculations. For example, if the difference in the photovoltaic connector mating gap exceeds 0.2mm, the difference in connector temperature rise exceeds 8℃, or the difference in the flatness of the junction box housing exceeds 0.5mm, or the RGB color shift exceeds 15%, then the corresponding component is determined to be abnormal, and photovoltaic connector status parameters and junction box status parameters are generated. In addition, a status diagnosis of the photovoltaic backsheet body is required. Deformation features refer to key indicators characterizing material degradation, such as the height of backsheet bulges, the width of delamination cracks, and the area of ​​yellowing regions. Image data of the backsheet region needs to be extracted from the image segmentation results, and then the deformation features are quantified and extracted using algorithms such as optical flow and deformation gradient operators to obtain the corresponding deformation parameters. Finally, these deformation parameters are weighted with the junction box status parameters and photovoltaic connector status parameters to generate target status parameters that include the status of electrical components and the backsheet material, thus completing the comprehensive status determination of the target location of the photovoltaic backsheet.

[0058] It is evident that by constructing benchmark samples, accurately extracting target areas, comparing multi-dimensional parameters, and fusing multiple indicators, quantitative diagnosis of the status of key components of photovoltaic backsheets has been achieved, eliminating the subjective errors of manual experience-based judgment and realizing layered diagnosis of electrical components and the backsheet body.

[0059] In one possible embodiment, the photovoltaic connector image parameters include: color feature parameters, texture feature parameters, and shape feature parameters. The step of determining the photovoltaic connector's state parameters based on the first photovoltaic connector image parameters and the second photovoltaic connector parameters, to obtain the photovoltaic connector state parameters, specifically includes the following steps: A41. Determine the first color feature parameter, the first texture feature parameter, and the first shape feature parameter based on the first photovoltaic connector image parameters; A42. Determine the second color feature parameter, the second texture feature parameter, and the second shape feature parameter based on the second photovoltaic connector image parameters; A43. Determine the difference between the first color feature parameter and the second color feature parameter to obtain the color feature difference; A44. Determine the difference between the first texture feature parameter and the second texture feature parameter to obtain the texture feature difference; A45. Determine the difference between the first shape feature parameter and the second shape feature parameter to obtain the deformation feature difference; A46. The color feature difference, texture feature difference, and deformation feature difference are weighted and calculated based on a preset difference parameter calculation formula to obtain a comprehensive difference parameter; A47. Based on the preset mapping relationship between the comprehensive difference parameters and the preset photovoltaic connector status parameters, determine the status parameters corresponding to the comprehensive difference parameters to obtain the photovoltaic connector status parameters.

[0060] Among them, color feature parameters are quantitative indicators used to characterize the color attributes of the connector shell, including RGB three-color component values, hue, saturation, etc.; texture feature parameters are parameters reflecting the surface roughness and wear marks of the connector, extracted through algorithms such as gray-level co-occurrence matrix and texture energy; shape feature parameters cover the geometric morphology indicators of the connector, such as the gap size at the mating point, the regularity of the shell contour, and the pin arrangement spacing. By decoupling the features of the first photovoltaic connector image parameters, the first color feature parameters, the first texture feature parameters, and the first shape feature parameters can be obtained respectively. Correspondingly, feature decoupling is performed on the second photovoltaic connector image parameters to determine the second color feature parameters, the second texture feature parameters, and the second shape feature parameters. Color feature difference is obtained by calculating the absolute difference between the corresponding indices of the first and second color feature parameters, such as RGB component difference and saturation difference. If the connector turns yellow or black due to aging, this difference will increase significantly. Texture feature difference is composed of the difference between the corresponding indices of two types of texture parameters, such as texture energy difference and roughness difference. Texture changes on the connector surface caused by wear and oxidation can be reflected by this difference. Shape feature difference is obtained by comparing the differences in geometric morphology indices, such as mating gap difference and contour regularity difference. Loosening of the connector and structural deformation caused by poor connection will cause this difference to exceed the normal range. The three types of differences reflect the deviation of the connector's state from three dimensions: appearance color, surface texture, and geometric structure. The preset difference parameter calculation formula is a weighted calculation model combining the weights of photovoltaic connector fault types. The color feature difference weight is set to 0.3, primarily reflecting aging anomalies; the texture feature difference weight is set to 0.2, focusing on wear-related issues; and the shape feature difference weight is set to 0.5, highlighting structural deformation and potential loose connections. The formula is: Comprehensive Difference Parameter = Color Feature Difference × 0.3 + Texture Feature Difference × 0.2 + Shape Feature Difference × 0.5. This weighted calculation yields a comprehensive difference parameter that reflects the degree of deviation from the connector's condition. The preset mapping relationship is constructed based on extensive experimental data. By statistically analyzing the range of comprehensive difference parameters corresponding to different fault states (such as minor loose connections, severe loose connections, mild aging, and severe aging), a one-to-one mapping table is established.

[0061] It is evident that by comparing and weighting multi-dimensional feature parameters, a quantitative assessment of the photovoltaic connector status is achieved. This avoids the one-sidedness of judging with a single feature parameter and highlights the impact of key faults such as structural deformation through weight allocation, thereby improving the accuracy of anomaly identification.

[0062] Step S360: If the first abnormal location is detected, wait for a preset time and then continue to perform data acquisition and abnormal detection operations on the first end face; and / or, send abnormal information generated in response to the abnormal detection operation to the cloud platform.

[0063] The first anomaly location is a potential risk point identified by the main processor during temperature screening, image segmentation, and parameter comparison. This risk point may be localized heating caused by poor soldering of the MC4 connector, localized temperature rise caused by poor contact inside the junction box or aging of the backplane material, or other abnormal thermal phenomena caused by environmental factors or component quality issues.

[0064] Specifically, when the main processor determines that the first abnormal location exists, it first enters a waiting phase. This waiting time can be adaptively set according to different abnormality types and data characteristics. For example, a shorter waiting time can be set for areas with large temperature fluctuations, while a longer time can be set for signals significantly affected by environmental noise, in order to filter out occasional factors such as light disturbances and temperature fluctuations caused by wind. After the preset time expires, the main processor resends the data acquisition command to the detection module, enabling the detection module to acquire the image and thermal imaging data of the first end face again. Through this repeated acquisition mechanism, the main processor can determine whether the suspected abnormal point is a persistent abnormality. Abnormal information can include the coordinates of the abnormal location, abnormality type, abnormality level, corresponding temperature value, image difference parameters, and the location's number in the component array. Then, the main processor sends the abnormal information to the cloud platform or backend server via the communication module. The cloud platform can connect to the power plant operation and maintenance system, fusing and analyzing the abnormal information with historical inspection data, component operating status, and environmental data, thereby providing operation and maintenance personnel with functions such as fault trend prediction, maintenance plan formulation, and equipment health analysis. In addition, for some serious anomalies that require manual intervention, such as MC4 connector overheating to the risk threshold, obvious damage to the junction box shell, or localized yellowing or blackening of the back panel, the main processor can directly trigger alarms through the cloud platform, enabling maintenance personnel to go to the site in a timely manner to handle the situation and reduce losses.

[0065] In one possible embodiment, sending the anomaly information generated in response to the anomaly detection operation to the cloud platform specifically includes the following steps: 361. Extract the data corresponding to the first abnormal position from the first data to obtain abnormal data; 362. Determine the anomaly type and anomaly level corresponding to the abnormal data; 363. Determine the anomaly information of the first anomaly location based on the anomaly type and anomaly level, and send the anomaly information to the cloud platform.

[0066] The anomaly types include electrical connection anomalies (e.g., loose MC4 connector connections, poor soldering of junction boxes) and material condition anomalies (e.g., blistering, yellowing, delamination, and cracking of the backsheet). These are determined by comparing characteristic parameters in the anomaly data with a preset fault feature library. For example, if a local temperature rise exceeds a threshold in the temperature data and is accompanied by an abnormal connector mating gap in the image, it is determined to be a loose MC4 connector connection. The anomaly level is divided into three levels based on the degree of impact on the photovoltaic system's operation: minor anomaly, moderate anomaly, and severe anomaly. A minor anomaly refers to parameters deviating slightly from the baseline range with no immediate safety risk, such as slight yellowing of the backsheet. A moderate anomaly refers to parameters deviating significantly and potentially affecting equipment lifespan, such as slight poor soldering of the junction box. A severe anomaly indicates an immediate safety hazard, such as severe loose MC4 connector connections accompanied by high temperature rise. The level is determined through a comprehensive evaluation of multiple dimensions, including the difference between the anomaly data and the baseline parameters, and the fault development trend. The abnormal information is sent through the remote communication module of the inspection robot. Based on wireless transmission, the abnormal information is uploaded to the cloud platform in real time. Data encryption is used during transmission to ensure information security. It also has the function of resuming transmission after interruption to avoid information loss due to network fluctuations.

[0067] It is evident that by extracting abnormal data and determining the type and level of abnormalities, the accuracy and completeness of information uploaded to the cloud platform are ensured; the standardized information structure and reliable transmission mechanism enable the abnormality handling process to be automated and traceable, helping operation and maintenance personnel to formulate efficient scheduling strategies based on big data.

[0068] Step S370: If the first abnormal position is not detected, the attitude adjustment command is sent to the rotating lifting mechanism, and / or the second walking command is sent to the walking mechanism.

[0069] If no anomalies are detected at this location, it indicates that the photovoltaic connectors, junction boxes, and backplanes within the current inspection area are all within normal range. In this case, the inspection robot needs to continue performing the inspection task along the preset route on the back of the photovoltaic modules. Therefore, the main processor needs to generate new posture adjustment instructions or second walking instructions based on the inspection path planning, the robot's current posture state, and the location of the target inspection area to ensure that the robot can complete the inspection of the entire photovoltaic backplane area without any omissions.

[0070] Specifically, when the main processor receives data from the detection module and confirms that the current area is not abnormal, it first determines whether the robot needs to adjust its posture. Since the backsheet of the photovoltaic module may have height variations along its length, different bracket obstruction positions, or slight deviations in the installation angle of the guide components, to ensure consistency and stability between the subsequently acquired images and thermal imaging data, the main processor generates new posture adjustment commands based on the angle, height, and horizontal offset data fed back by the posture detection module. These posture adjustment commands can include rotation commands, lifting commands, or a combination of rotation and lifting commands to correct the posture of the detection module, maintaining a suitable shooting angle and distance from the photovoltaic backsheet. If the current posture meets the requirements for subsequent inspection, the main processor generates a second walking command based on the inspection path model. The goal of the second walking command is to drive the walking mechanism to continue moving along the guide components to the next inspection area. The command typically includes parameters such as walking direction, walking distance, speed threshold, and deceleration point position. For example, when the robot approaches a new inspection area, the walking mechanism automatically reduces its walking speed for precise positioning and posture calibration. In addition, the main processor monitors the operational status of the walking mechanism in real time, such as whether the wheelsets maintain stable contact with the guides and whether the curved pressure plates provide sufficient clamping, to ensure the robot operates smoothly in complex rooftop environments. During the execution of the second walking command, the main processor also continuously monitors parameters such as ambient temperature, robot tilt angle, and walking resistance to determine whether dynamic adjustments to the commands are needed based on environmental changes. For example, in the event of strong winds, support vibration, or local tilting of the guides, the main processor can pause the walking mechanism and resend the attitude adjustment command to ensure the robot maintains high stability at critical positions.

[0071] It is evident that by establishing automatic detection under normal area conditions, the inspection robot can achieve continuous and complete inspection of photovoltaic backsheets. The posture adjustment command ensures that the detection module always maintains a complete shooting angle, improving image clarity and thermal imaging accuracy. The second walking command enables the inspection robot to move smoothly along the guide to the next inspection area, avoiding inspection interruptions and missed inspections caused by manual intervention or discontinuous paths.

[0072] In one possible embodiment, sending the attitude adjustment command to the rotary lifting mechanism and / or sending the second walking command to the walking mechanism specifically includes the following steps: 371. When the walking mechanism is controlled to travel to the target position, the second data is obtained through the detection module; the target position is the area near the junction box or photovoltaic connector of the guide member; 372. Based on the attitude adjustment command, the rotary lifting mechanism obtains a rotation command and / or a lifting command; 373. Control the rotating and lifting mechanism to rotate and / or lift according to the rotation, so as to drive the detection module to collect data on the first end face.

[0073] The target position guide rail or wire rope includes the area near the junction box or photovoltaic connector (MC4 connector, junction box); the second data refers to the raw data of the target area initially collected by the detection module after the walking mechanism reaches the target position, used for subsequent fine-tuning of the mechanism's posture and for the processor to detect abnormal positions based on the second data. Specifically, the main controller drives the walking mechanism to move along the guide according to the preset path instructions, and triggers the detection module to collect data when it reaches the target position. The posture adjustment instruction is the mechanism action instruction generated by the main processor based on the second data and preset acquisition standards, which aims to eliminate occlusion interference and optimize the acquisition angle by adjusting the spatial posture of the detection module; the rotation instruction is used to control the rotating lifting mechanism to achieve 360° rotation in the horizontal direction, and the lifting instruction is used to control the mechanism to move up and down in the vertical direction. The two types of instructions can be issued individually or in combination, depending on the spatial position characteristics of the target area.

[0074] Specifically, the main controller analyzes the second data. If there is an angular deviation between the detection module and the junction box / photovoltaic connector, causing partial shading, a rotation command is generated. If the acquisition distance is too far or too close, affecting data clarity, a lifting command is generated. If both angular and distance issues exist simultaneously, a combination of rotation and lifting commands is generated synchronously to ensure that the commands accurately match the actual acquisition requirements. The rotation and lifting mechanism is a mechanical structure with multi-degree-of-freedom adjustment capabilities. It is fixedly connected to the detection module and can flexibly change the spatial position and angle of the detection module through posture adjustment. The first end face is the detection surface of the photovoltaic backsheet, which includes key detection areas such as the junction box, photovoltaic connector, and backsheet body.

[0075] It is evident that by moving the walking mechanism to the target position and adjusting the posture through the multi-degree-of-freedom rotation and lifting mechanism, the blind spots of key components of the photovoltaic backsheet are effectively reduced, and data acquisition of the junction box and photovoltaic connector area is realized. Based on the coordinated linkage between the dual-light detection module and the mechanism's movements, the accuracy and completeness of the collected data are ensured, the risk of missed detection due to poor acquisition angle is reduced, and the intelligence and efficiency of photovoltaic backsheet inspection are further enhanced.

[0076] In one possible embodiment, the first data includes a first image and first heatmap data; the detection module includes an infrared camera, a visible light camera, and a communication module; receiving the first data specifically includes the following steps: B1. Receive the first data acquisition command sent by the main processor; B2. Determine the image acquisition command and the heat map acquisition command according to the first data acquisition command; B3. Acquire a first image of the first end face using the visible light camera in response to the image acquisition command, and acquire first thermal imaging data of the first end face using the infrared camera according to the thermal map acquisition command; B4. Send the first thermal imaging data and the first image to the main processor through the communication module.

[0077] The detection module receives the first data acquisition command and analyzes it using its built-in command parsing module. The image acquisition command, generated for the visible light camera, includes parameters such as image resolution, frame rate, and focus mode, used to acquire information about the appearance of the photovoltaic backsheet. The thermal map acquisition command, customized for the infrared camera, includes settings such as temperature range, temperature resolution, and thermal imaging mode, used to capture temperature distribution data of the target area. The command parsing process matches the hardware parameters of the two types of cameras to ensure that the generated acquisition commands are compatible with the equipment performance, achieving synchronized dual-light acquisition and preventing data failure due to mismatched command parameters. The communication module is the data transmission unit of the detection module, supporting remote wireless communication. Its transmission technology features strong anti-interference capabilities and stable transmission rates, adapting to the complex electromagnetic environment of photovoltaic power plants. After the detection module completes the first thermal imaging data and first image acquisition, the communication module performs format standardization and compression encoding on both types of data to reduce data transmission volume and ensure data integrity, then sends the data to the main processor via a wireless transmission link. Thus, through the collaborative acquisition of data by dual-light cameras and reliable transmission via the communication module, simultaneous acquisition of data on the appearance and temperature distribution of the photovoltaic backsheet was achieved. This not only covered the monitoring of appearance defects but also detected temperature anomalies, overcoming the limitations of single acquisition methods. The parsing of the first acquisition command and the coordinated control of the equipment ensured the spatiotemporal consistency and accuracy of the acquired data, providing a high-quality data foundation for subsequent intelligent algorithm analysis and facilitating real-time and precise monitoring of the photovoltaic backsheet's condition.

[0078] In one possible embodiment, the walking mechanism includes a set of walking wheels and a drive mechanism disposed along the guide member; the set of walking wheels includes at least one walking wheel, the at least one walking wheel having a groove in its circumference, the groove cooperating with the guide member, the groove being used to guide the movement of the walking mechanism; the second end face of the set of walking wheels is provided with an arc-shaped pressure plate, the arc-shaped pressure plate forming a clamping structure around the guide member, the clamping structure being used to stably support the walking mechanism on the guide member.

[0079] In the design of the walking mechanism of the photovoltaic backsheet inspection robot, the coordinated operation of the walking wheel set and the drive mechanism, as well as the design of the clamping structure, are crucial to ensuring the robot's stable movement along the guide. The walking mechanism, as the motion execution unit of the inspection robot, includes a walking wheel set and a drive mechanism positioned along the guide (a steel wire rope or simple guide rail fixed to the photovoltaic backsheet). The drive mechanism provides power output to the walking wheel set and can employ a combination of a micro motor and a reduction gear set. The output speed and torque are adjusted through a control module to achieve forward, backward, and start / stop control of the walking mechanism, ensuring that the movement speed matches the inspection data acquisition. The walking wheel set contains at least one walking wheel, and the circumferentially grooved groove is a key structure for achieving the guiding function. The cross-sectional shape and size of the groove match the type of guide. If the guide component is a steel wire rope, the groove is designed as an arc-shaped concave surface with a radius of curvature matching the diameter of the steel wire rope, ensuring a tight fit between the groove and the surface of the steel wire rope. If the guide component is a simple guide rail, the groove is designed as a rectangular or trapezoidal structure matching the cross-section of the guide rail. Through the interlocking fit between the groove and the guide component, the horizontal deviation of the walking mechanism is limited, achieving precise guidance in the direction of movement and preventing the walking mechanism from deviating from the preset path due to the installation tilt angle or vibration of the photovoltaic backsheet. The arc-shaped pressure plate on the second end face of the walking wheel assembly (the end face opposite to the drive mechanism) is a core auxiliary structure for improving the stability of the walking mechanism. The arc-shaped pressure plate is made of elastic metal material, and its curvature is consistent with the surface curvature of the guide component. It is fixed to the walking wheel assembly body by bolts or elastic connectors, forming a semi-enclosed clamping structure around the guide component. The working principle of this clamping structure is as follows: the arc-shaped pressure plate generates a continuous clamping force through its own elastic deformation, keeping the groove of the walking wheel assembly and the guide component in a tight fit at all times, while offsetting the vertical vibration and lateral force experienced by the walking mechanism during movement. For example, when the inspection robot travels to an inclined area of ​​the photovoltaic backsheet or encounters airflow disturbances, the clamping structure can effectively prevent the walking wheels from detaching from the guide, ensuring stable support of the walking mechanism on the guide and avoiding data ambiguity or robot shutdown due to bumps. Furthermore, the elastic design of the arc-shaped pressure plate also provides a certain degree of cushioning, reducing frictional wear between the walking wheels and the guide, extending the equipment's service life, and accommodating slight dimensional deviations on the guide surface, thus improving structural adaptability.

[0080] As can be seen, the walking mechanism achieves directional guidance through the precise engagement of the grooves and guide components, and the clamping structure formed by the arc-shaped pressure plate ensures stable support, effectively solving the problem of easy deviation and detachment of the walking mechanism in the photovoltaic backsheet inspection scenario. In addition, the power output design of the drive mechanism can flexibly adapt to the inspection rhythm, and the structural adaptability of the walking wheel set improves the compatibility of the mechanism with different types of guide components. The overall structural design takes into account both motion accuracy and operational stability, providing reliable motion guarantee for the inspection robot to achieve blind-spot-free data collection, while extending the service life of the equipment and reducing maintenance costs.

[0081] For easier understanding, please refer to Figure 5 , Figure 5 This is a schematic diagram of a walking module provided in an embodiment of this application. As can be seen, the walking module 500 includes an arc-shaped pressure plate, a drive mechanism, and a pair of walking wheels. The drive mechanism drives the walking wheels to move along a steel wire rope, enabling the inspection robot to achieve stable and controllable forward and backward movement along the steel wire rope within the narrow space of the photovoltaic backsheet, thus completing linear movement along the inspection path. Specifically, the steel wire rope is laid below the photovoltaic backsheet and serves as a guide for the inspection robot. The walking wheels have circumferential grooves that match the steel wire rope, allowing the walking wheels to fit tightly against the steel wire rope for stable contact support. Driven by the drive mechanism, the walking wheels roll around the steel wire rope. The drive mechanism may include a micro motor, a reduction mechanism, and a transmission gear set, enabling the walking wheels to operate smoothly under low-speed, high-torque driving conditions, avoiding the impact of shaking or slippage on the quality of inspection data acquisition. The arc-shaped pressure plate is located on the outer side of the walking wheels, and its arc shape matches the outer diameter of the steel wire rope, forming a clamping structure together with the walking wheels. This clamping structure provides a circumferential restraint on the wire rope, ensuring reliable support for the walking mechanism even during high-altitude operations or tilted installations of photovoltaic modules. The arc-shaped pressure plate can be connected to the walking mechanism body via elastic elements or threaded fasteners, allowing for flexible adjustment to accommodate minor deformations of the wire rope caused by temperature changes or tension adjustments. Furthermore, the drive mechanism is typically mounted on the main frame of the walking mechanism and connected to the wheels via a drive shaft. Under the control of the first or second walking command issued by the main processor, the drive mechanism achieves real-time adjustment of the wheel speed and direction, ensuring the inspection robot can accurately stop at key locations such as photovoltaic connectors, junction boxes, or backplanes according to the data collection requirements of the inspection module.

[0082] It is evident that the design of this walking mechanism enables the inspection robot to achieve highly stable and reliable linear motion under the guidance of the steel wire rope, providing an important mechanical foundation for the precise positioning and continuous inspection of the inspection module, and significantly improving the adaptability and operational safety of photovoltaic backsheet inspection in complex environments.

[0083] For easier understanding, please refer to Figure 6 , Figure 6 This is a schematic diagram of a scenario where an inspection robot inspects a photovoltaic backsheet, as provided in an embodiment of this application. As can be seen in the schematic diagram 600, the backsheet inspection robot is installed on the back of the photovoltaic backsheet, that is, on the other side of the photovoltaic backsheet that is exposed to sunlight, in order to inspect the MC4 connectors and junction boxes installed on the photovoltaic backsheet, or to inspect the condition of the photovoltaic backsheet, so as to check whether these components in the photovoltaic backsheet have any safety hazards.

[0084] Specifically, the back panel inspection robot moves on a pair of parallel steel cables fixed to the brackets of the photovoltaic back panels. This allows the robot to inspect narrow areas. The steel cable track can adapt to the narrow and inclined space behind the photovoltaic back panels in the photovoltaic power station, effectively solving the limitation of traditional inspection equipment on operating space. This enables the back panel inspection robot to move along the steel cable track in different areas on the back of the photovoltaic back panels, covering the distribution points of MC4 connectors, junction boxes, and key areas of the photovoltaic back panels. The photovoltaic back panels are arranged in an array, and the steel cable track extends along the longitudinal or lateral direction of the photovoltaic back panels. The back panel inspection robot moves along the track under the action of the drive mechanism by engaging its walking mechanism with the steel cables. At the same time, it collects images and thermal data of the photovoltaic back panels and related components based on the onboard detection module (such as a dual-light camera), providing data support for subsequent status analysis and anomaly identification. The overall scenario fully demonstrates the deployment logic and operation mode of the inspection robot on the back of the photovoltaic back panels.

[0085] For easier understanding, please refer to Figure 7 , Figure 7 This is a schematic diagram of the installation structure of an inspection robot on a photovoltaic backsheet according to an embodiment of this application. As shown in the schematic diagram 700, the inspection robot is mounted on a steel wire rope, with both ends of the rope fixed to a photovoltaic backsheet support, which also serves to secure the photovoltaic backsheet. Specifically, two parallel steel wire rope tracks are fixed to the crossbeam structure of the photovoltaic backsheet support via connectors. The extension direction of the tracks is consistent with the array arrangement direction of the photovoltaic backsheet, covering the longitudinal or transverse areas of the photovoltaic backsheet and providing a full-stroke movement path for the inspection robot. The circumferential grooves of the robot's wheels engage with the surface of the steel wire rope, and the clamping structure formed by the arc-shaped pressure plate ensures stable attachment of the robot to the steel wire rope. Even in scenarios where the photovoltaic backsheet is deployed at an angle, the robot maintains structural stability during movement, avoiding the risk of detachment due to vibration or tilt. The backsheet inspection robot is positioned on the back of the photovoltaic backsheet (i.e., the non-sunlit side). Its onboard inspection modules (such as dual-light cameras) can directly face the photovoltaic backsheet body, MC4 connectors, junction boxes, and other target components, ensuring that the viewing angle and distance for data collection meet the inspection accuracy requirements. Furthermore, the fixed points of the wire rope track and the connection method between the robot and the track all use standardized interfaces, facilitating adaptation and adjustment to different specifications of photovoltaic backsheet brackets, while reducing the complexity of on-site installation and maintenance.

[0086] As can be seen, by implementing the method described in the above embodiments of the photovoltaic backsheet inspection method, the potential problems of the photovoltaic backsheet can be accurately identified, thereby improving the efficiency of photovoltaic backsheet inspection.

[0087] The above primarily describes the solutions of the embodiments of this application from the perspective of the method execution process. It is understood that, in order to achieve the above functions, the robot includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments provided herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0088] This application embodiment can divide the robot into functional units according to the above method example. For example, each function can be divided into a separate functional unit, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0089] When dividing each function into modules according to its corresponding function. Figure 8 This is a block diagram of an inspection robot for photovoltaic backsheets provided in an embodiment of this application. The inspection robot includes: a main processor, a walking mechanism, a rotating and lifting mechanism, a detection module mounted on the rotating and lifting mechanism, and a guide component mounted on a first end face of the photovoltaic backsheet. The guide component is at least one of a guide rail or a steel wire rope. The first end face is the opposite side of the photovoltaic backsheet facing the light source. The inspection robot 800 for photovoltaic backsheets includes: The control processing unit 810 is used to receive and respond to the first walking command for the inspection robot through the main processor, send the first posture adjustment command to the rotating lifting mechanism, and send the first data acquisition command to the detection module. The walking adjustment unit 820 is used to receive the first walking command through the walking mechanism and, in response to the first walking command, perform forward and / or backward movement on the guide member; The attitude adjustment unit 830 is used to receive and respond to the attitude adjustment command through the rotary lifting mechanism; The data acquisition unit 840 is configured to receive and respond to the first data acquisition command through the detection module; acquire first data from the first end face; and send the first data to the main processor. The control processing unit 810 is further configured to receive the first data, detect whether there is a first abnormal position on the first end face based on the first data; if the first abnormal position is detected, wait for a preset time and then continue to perform data acquisition and abnormal detection operations on the first end face; and / or send abnormal information generated in response to the abnormal detection operation to the cloud platform; if the first abnormal position is not detected, send the attitude adjustment command to the rotating lifting mechanism, and / or send a second walking command to the walking mechanism.

[0090] In one possible embodiment, the first data includes first image data and first environmental data, and the control processing unit 810, in detecting whether a first abnormal position exists on the first end face based on the first data, is specifically configured to: The temperatures corresponding to multiple locations in the first end face are extracted from the first environmental data to obtain multiple temperatures; each of the multiple temperatures corresponds one-to-one with each of the multiple locations. When the target temperature is greater than a preset temperature threshold, the image of the target location corresponding to the target temperature is selected from the first image data to obtain the target image; the target temperature is any one of the plurality of temperatures; Based on a preset image segmentation algorithm, the target location in the target image is segmented to obtain the image segmentation result; Based on the image segmentation results, the state parameters of the target location are determined to obtain the target state parameters; If the target state parameter indicates an abnormal state, the target position is recorded as the first abnormal position; If the target status parameter indicates that the status is normal, then the target status parameter is sent to the cloud platform.

[0091] In one possible embodiment, the control processing unit 810, in determining the state parameters of the target location based on the image segmentation result and obtaining the target state parameters, is specifically configured to: Acquire images of a normal photovoltaic connector and a normal junction box; The image segmentation results for the photovoltaic connector and the junction box are extracted from the image segmentation results to obtain the photovoltaic connector image and the junction box image; Determine the image parameters corresponding to the photovoltaic connector image and the junction box image to obtain the first photovoltaic connector image parameters and the first junction box image parameters; Determine the image parameters corresponding to the normal photovoltaic connector image and the normal junction box image to obtain the second photovoltaic connector image parameters and the second junction box image parameters; The photovoltaic connector status parameters are determined based on the first photovoltaic connector image parameters and the second photovoltaic connector image parameters to obtain the photovoltaic connector status parameters. The state parameters of the junction box are determined based on the first junction box image parameters and the second junction box image parameters to obtain the junction box state parameters. Image data of the back plate in the first end face is extracted from the image segmentation result to obtain the back plate image; Deformation features are extracted from the backplate image to obtain deformation parameters; The target state parameters are determined based on the deformation parameters, the junction box state parameters, and the photovoltaic connector state parameters.

[0092] In one possible embodiment, the photovoltaic connector image parameters include: color feature parameters, texture feature parameters, and shape feature parameters. The control processing unit 810, in determining the photovoltaic connector state parameters based on the first photovoltaic connector image parameters and the second photovoltaic connector parameters, specifically performs the following: The first color feature parameter, the first texture feature parameter, and the first shape feature parameter are determined based on the first photovoltaic connector image parameters; The second color feature parameter, the second texture feature parameter, and the second shape feature parameter are determined based on the image parameters of the second photovoltaic connector. The difference between the first color feature parameter and the second color feature parameter is determined to obtain the color feature difference; The difference between the first texture feature parameter and the second texture feature parameter is determined to obtain the texture feature difference; The difference between the first shape feature parameter and the second shape feature parameter is determined to obtain the deformation feature difference; The color feature difference, texture feature difference, and deformation feature difference are weighted and calculated based on a preset difference parameter calculation formula to obtain a comprehensive difference parameter. Based on the mapping relationship between the preset comprehensive difference parameters and the preset photovoltaic connector status parameters, the status parameters corresponding to the comprehensive difference parameters are determined, and the photovoltaic connector status parameters are obtained.

[0093] In one possible embodiment, the control processing unit 810, in sending the anomaly information generated in response to the anomaly detection operation to the cloud platform, is specifically configured to: Extract the data corresponding to the first abnormal position from the first data to obtain abnormal data; Determine the anomaly type and anomaly level corresponding to the abnormal data; The anomaly information of the first anomaly location is determined according to the anomaly type and anomaly level, and the anomaly information is sent to the cloud platform.

[0094] In one possible embodiment, the control processing unit 810, in sending the attitude adjustment command to the rotary lifting mechanism and / or sending the second walking command to the walking mechanism, is specifically configured to: When the walking mechanism is controlled to travel to the target position, the detection module acquires second data; the target position is the area near the junction box or photovoltaic connector of the guide member; The rotation and lifting mechanism is determined to receive rotation commands and / or lifting commands based on the attitude adjustment commands. The rotating and lifting mechanism is controlled to rotate and / or lift according to the rotation, so as to drive the detection module to collect data on the first end face.

[0095] In one possible embodiment, the first data includes a first image and first heatmap data; the detection module includes an infrared camera, a visible light camera, and a communication module; the data acquisition unit 840 is specifically used for: Receive the first data acquisition command sent by the main processor; The image acquisition command and the heat map acquisition command are determined according to the first data acquisition command; The visible light camera acquires a first image of the first end face in response to the image acquisition command, and the infrared camera acquires first thermal imaging data of the first end face according to the thermal map acquisition command. The first thermal imaging data and the first image are sent to the main processor through the communication module.

[0096] As can be seen, this application provides an inspection robot for photovoltaic backsheets. The inspection robot includes: a main processor, a walking mechanism, a rotating and lifting mechanism, a detection module mounted on the rotating and lifting mechanism, and a guide component mounted on the first end face of the photovoltaic backsheet. The guide component is at least one of a guide rail or a steel wire rope. The first end face is the opposite side of the photovoltaic backsheet facing the light source. The main processor is used to receive and respond to a first walking command for the inspection robot, send a first posture adjustment command to the rotating and lifting mechanism, and send a first data acquisition command to the detection module. The walking mechanism is used to receive the first walking command and respond to the first walking command by moving forward and / or backward on the guide component. A rotating lifting mechanism is used to receive and respond to attitude adjustment commands; a detection module is used to receive and respond to a first data acquisition command; acquire first data from the first end face; and send the first data to the main processor; the main processor is used to receive the first data, detect whether there is a first abnormal position on the first end face based on the first data; if the first abnormal position is detected, it waits for a preset time and then continues to perform data acquisition and abnormal detection operations on the first end face; and / or sends abnormal information generated in response to the abnormal detection operation to the cloud platform; if the first abnormal position is not detected, it sends the attitude adjustment command to the rotating lifting mechanism, and / or sends a second walking command to the walking mechanism. This improves inspection efficiency and the accuracy of identifying potential problems with photovoltaic backsheets.

[0097] This application also provides an inspection system for photovoltaic backsheets, wherein the inspection system for photovoltaic backsheets can perform some or all of the steps of any of the methods described in the above method embodiments, and the computer includes a robot.

[0098] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments. The computer program product may be a software installation package, and the computer may include a robot.

[0099] It should be noted that, for the sake of simplicity, the above embodiments are all described as a series of actions. Those skilled in the art should understand that this application is not limited to the described order of actions, as some steps in the embodiments of this application can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions, steps, modules, or units involved are not necessarily essential to the embodiments of this application.

[0100] In the above embodiments, the descriptions of each embodiment in this application have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0101] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

[0102] The steps of the methods or algorithms described in the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in RAM, flash memory, ROM, EPROM, electrically erasable programmable read-only memory (EEPROM), registers, hard disk, portable hard disk, read-only optical disk (CD-ROM), or any other form of storage medium well known in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Furthermore, the ASIC can reside in a terminal device or management device. Alternatively, the processor and storage medium can exist as discrete components in the terminal device or management device.

[0103] The modules / units included in the various devices and products described in the above embodiments can be software modules / units, hardware modules / units, or a combination of both. For example, for devices and products applied to or integrated into a chip, all modules / units can be implemented using hardware methods such as circuits, or at least some modules / units can be implemented using software programs that run on a processor integrated within the chip, while the remaining (if any) modules / units can be implemented using hardware methods such as circuits. For devices and products applied to or integrated into a chip module, all modules / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components of the chip module, or at least some modules / units can be implemented using hardware methods such as circuits. The implementation is achieved through a software program that runs on the processor integrated within the chip module. The remaining modules / units (if any) can be implemented using hardware methods such as circuits. For various devices and products applied to or integrated into terminal equipment, each of their modules / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components within the terminal equipment. Alternatively, at least some modules / units can be implemented through a software program that runs on the processor integrated within the terminal equipment, while the remaining modules / units (if any) can be implemented using hardware methods such as circuits.

[0104] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the embodiments of this application. It should be understood that the above descriptions are merely specific embodiments of the embodiments of this application and are not intended to limit the protection scope of the embodiments of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solutions of the embodiments of this application should be included within the protection scope of the embodiments of this application.

Claims

1. An inspection robot for photovoltaic backsheets, characterized in that, The inspection robot includes: a main processor, a walking mechanism, a rotating and lifting mechanism, a detection module mounted on the rotating and lifting mechanism, and a guide component mounted on the first end face of the photovoltaic backsheet. The guide component is at least one of a guide rail or a steel wire rope. The first end face is the opposite side of the photovoltaic backsheet facing the light-receiving surface. The main processor is used to receive and respond to the first walking command for the inspection robot, send the first posture adjustment command to the rotating lifting mechanism, and send the first data acquisition command to the detection module. The walking mechanism is used to receive the first walking command and respond to the first walking command by moving forward or backward on the guide member. The rotary lifting mechanism is used to receive and respond to the attitude adjustment command; The detection module is used to receive and respond to the first data acquisition command; acquire first data from the first end face; and send the first data to the main processor. The main processor is configured to receive the first data, which includes first image data and first environmental data; extract temperatures corresponding to multiple locations on the first end face from the first environmental data to obtain multiple temperatures; each of the multiple temperatures corresponds one-to-one with each of the multiple locations; when the target temperature is greater than a preset temperature threshold, filter the image of the target location corresponding to the target temperature from the first image data to obtain a target image; the target temperature is any one of the multiple temperatures; perform image segmentation on the target location in the target image based on a preset image segmentation algorithm to obtain an image segmentation result; determine the state parameters of the target location based on the image segmentation result to obtain target state parameters; if the target state parameters indicate an abnormal state, record the target location as a first abnormal location; if the target state parameters indicate a normal state, send the target state parameters to the cloud platform. If the first abnormal location is detected, wait for a preset time and then continue to perform data acquisition and abnormal detection operations on the first end face; and / or send abnormal information generated in response to the abnormal detection operation to the cloud platform. If the first abnormal position is not detected, the attitude adjustment command is sent to the rotating lifting mechanism, and / or the second walking command is sent to the walking mechanism.

2. The inspection robot for photovoltaic backsheets as described in claim 1, characterized in that, In the process of determining the state parameters of the target location based on the image segmentation result and obtaining the target state parameters, the main processor is specifically used for: Acquire images of a normal photovoltaic connector and a normal junction box; The image segmentation results for the photovoltaic connector and the junction box are extracted from the image segmentation results to obtain the photovoltaic connector image and the junction box image; Determine the image parameters corresponding to the photovoltaic connector image and the junction box image to obtain the first photovoltaic connector image parameters and the first junction box image parameters; Determine the image parameters corresponding to the normal photovoltaic connector image and the normal junction box image to obtain the second photovoltaic connector image parameters and the second junction box image parameters; The photovoltaic connector status parameters are determined based on the first photovoltaic connector image parameters and the second photovoltaic connector image parameters to obtain the photovoltaic connector status parameters. The state parameters of the junction box are determined based on the first junction box image parameters and the second junction box image parameters to obtain the junction box state parameters. The image data of the back plate in the first end face is extracted from the image segmentation result to obtain the back plate image; Deformation features are extracted from the backplate image to obtain deformation parameters; The target state parameters are determined based on the deformation parameters, the junction box state parameters, and the photovoltaic connector state parameters.

3. The inspection robot for photovoltaic backsheets as described in claim 2, characterized in that, The photovoltaic connector image parameters include: color feature parameters, texture feature parameters, and shape feature parameters. In the process of determining the photovoltaic connector's state parameters based on the first and second photovoltaic connector image parameters, the main processor is specifically used for: The first color feature parameter, the first texture feature parameter, and the first shape feature parameter are determined based on the first photovoltaic connector image parameters; The second color feature parameter, the second texture feature parameter, and the second shape feature parameter are determined based on the second photovoltaic connector image parameters; The difference between the first color feature parameter and the second color feature parameter is determined to obtain the color feature difference; The difference between the first texture feature parameter and the second texture feature parameter is determined to obtain the texture feature difference; The difference between the first shape feature parameter and the second shape feature parameter is determined to obtain the deformation feature difference; The color feature difference, texture feature difference, and deformation feature difference are weighted and calculated based on a preset difference parameter calculation formula to obtain a comprehensive difference parameter. Based on the mapping relationship between the preset comprehensive difference parameters and the preset photovoltaic connector status parameters, the status parameters corresponding to the comprehensive difference parameters are determined, and the photovoltaic connector status parameters are obtained.

4. The inspection robot for photovoltaic backsheets as described in claim 1, characterized in that, During the process of sending the exception information generated by the exception detection operation to the cloud platform, the main processor is specifically used for: Extract the data corresponding to the first abnormal position from the first data to obtain abnormal data; Determine the anomaly type and anomaly level corresponding to the abnormal data; The anomaly information of the first anomaly location is determined according to the anomaly type and anomaly level, and the anomaly information is sent to the cloud platform.

5. The inspection robot for photovoltaic backsheets as described in claim 1, characterized in that, During the process of sending the attitude adjustment command to the rotary lifting mechanism and / or sending the second walking command to the walking mechanism, the main processor is specifically used for: When the walking mechanism is controlled to travel to the target position, the second data is obtained through the detection module; the target position is the area near the junction box or photovoltaic connector of the guide member; The rotation and lifting mechanism is determined to receive rotation commands and / or lifting commands based on the attitude adjustment commands. The rotating and lifting mechanism is controlled according to the rotation command and / or the lifting command to drive the detection module to collect data from the first end face.

6. The inspection robot for photovoltaic backsheets as described in claim 1, characterized in that, The first data includes a first image and a first heatmap; the detection module includes an infrared camera, a visible light camera, and a communication module; during the process of receiving the first data, the detection module is specifically used for: Receive the first data acquisition command sent by the main processor; The image acquisition command and the heat map acquisition command are determined according to the first data acquisition command; The visible light camera acquires a first image of the first end face in response to the image acquisition command, and the infrared camera acquires first thermal imaging data of the first end face according to the thermal map acquisition command. The first thermal imaging data and the first image are sent to the main processor through the communication module.

7. The inspection robot for photovoltaic backsheets as described in claim 1, characterized in that, The walking mechanism includes a set of walking wheels and a drive mechanism arranged along the guide member; the set of walking wheels includes at least one walking wheel, the at least one walking wheel has a groove in its circumference, the groove cooperates with the guide member, and the groove is used to guide the movement of the walking mechanism; the second end face of the set of walking wheels is provided with an arc-shaped pressure plate, the arc-shaped pressure plate forms a clamping structure around the guide member, and the clamping structure makes the walking mechanism stably supported on the guide member.

8. A method for inspecting photovoltaic backsheets, characterized in that, A main processor is applied to an inspection robot, the inspection robot comprising: a walking mechanism, a rotating and lifting mechanism, a detection module disposed on the rotating and lifting mechanism, and a guide component disposed on a first end face of a photovoltaic backsheet, the guide component being at least one of a guide rail or a steel wire rope; the first end face is the opposite side of the photovoltaic backsheet facing the light-receiving surface, the method comprising: Receive and respond to the first walking command for the inspection robot; Send a first attitude adjustment command to the rotary lifting mechanism; and send a first data acquisition command to the detection module; The detection module is controlled to receive and respond to the first data acquisition command; Collect first data from the first end face; and send the first data to the main processor; The system receives the first data, which includes first image data and first environmental data; it extracts temperatures corresponding to multiple locations on the first end face from the first environmental data to obtain multiple temperatures; each of the multiple temperatures corresponds one-to-one with each of the multiple locations; when the target temperature is greater than a preset temperature threshold, it filters the image of the target location corresponding to the target temperature from the first image data to obtain a target image; the target temperature is any one of the multiple temperatures; it performs image segmentation on the target location in the target image based on a preset image segmentation algorithm to obtain an image segmentation result; it determines the state parameter of the target location based on the image segmentation result to obtain a target state parameter; if the target state parameter indicates an abnormal state, it records the target location as a first abnormal location; if the target state parameter indicates a normal state, it sends the target state parameter to the cloud platform. If the first abnormal location is detected, wait for a preset time and then continue to perform data acquisition and abnormal detection operations on the first end face; and / or send abnormal information generated in response to the abnormal detection operation to the cloud platform. If the first abnormal position is not detected, the attitude adjustment command is sent to the rotating lifting mechanism, and / or the second walking command is sent to the walking mechanism.

9. An inspection system for photovoltaic backsheets, characterized in that, The inspection system for photovoltaic backsheets performs the method described in claim 8.

Images