A real-time image processing and secure transmission method for power unmanned aerial vehicle inspection

Abstract

The application discloses a real-time image processing and safe transmission method for power unmanned aerial vehicle inspection, and comprises the following steps: S1, collecting unmanned aerial vehicle inspection video streams in real time through a multi-path video access interface; S2, calling a hardware decoder to decode the video streams in real time, and displaying the video streams on a rugged terminal through a zero-copy rendering pipeline with low delay; S3, performing space-time alignment and fusion superposition on the real-time video, preloaded GIS geographic information and historical inspection data; S4, in response to touch or voice instructions of an operator, performing frame selection and labeling on discovered equipment defects on real-time video or playback video, and associating defect type and grade information; S5, based on a national secret SM4 algorithm, performing real-time encryption transmission on video streams, labeling data and control instructions in the inspection process; S6, after the inspection task is completed, generating a standardized inspection report containing defect images, position descriptions and processing suggestions through one key; and the technical problem of insufficient real-time performance of the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) power line inspection technology, and in particular to a real-time image processing and secure transmission method for UAV power line inspection. Background Technology

[0002] Currently, drones are widely used in power distribution network inspection operations, but existing inspection systems have the following technical shortcomings in practical applications: The video stream uses offline storage and post-event analysis, making real-time processing and assessment impossible on-site, resulting in poor timeliness of defect handling. Ordinary terminals have weak decoding capabilities and software decoding has high latency, making it difficult to support real-time display of 4K high-definition video. Video footage, GIS geographic information, and historical inspection data lack spatiotemporal alignment and fusion display, leading to low on-site decision-making efficiency. Defect annotation relies on manual post-event processing, which is cumbersome, prone to omissions, and lacks standardization. Inspection videos, annotation data, and control commands are not transmitted using compliant power encryption methods, posing risks of data leakage, tampering, and hijacking. Inspection reports need to be manually compiled, which is time-consuming, has low standardization, and cannot achieve rapid output of closed-loop tasks.

[0003] Existing technologies fail to integrate real-time image processing, multi-source fusion, on-site interactive annotation, national cryptographic security transmission, and automatic report generation into a unified process, thus failing to meet the actual operational needs of low latency, high reliability, strong security, and portability in power field operations. Summary of the Invention: To address the aforementioned technical issues, a real-time image processing and secure transmission method and system for power plant drone inspection is provided. This system enables seamless integration of real-time video acquisition, low-latency display, multi-source fusion, on-site annotation, encrypted transmission, and one-click report generation, resolving technical problems such as insufficient real-time performance, weak on-site processing capabilities, poor data fusion, non-compliance with security regulations, and low operational efficiency.

[0004] To address the aforementioned technical problems, the present invention provides a technical solution: a real-time image processing and secure transmission method for power plant drone inspection, characterized by comprising the following steps: S1: Synchronously acquire real-time visible light and infrared video streams from UAV inspections through multiple video access interfaces, establish timestamps and spatial coordinate benchmarks for video data, and provide raw data sources for subsequent real-time processing and spatiotemporal fusion. S2: Based on the acquired video stream, the hardware decoder is called to complete real-time decoding, and the decoded image is directly mapped to the ruggedized terminal for low-latency display through the zero-copy rendering pipeline, providing real-time image support for on-site observation and defect identification. S3: Based on the real-time video frame timestamps and UAV positioning information output by S2, the real-time video footage is aligned spatiotemporally with the pre-loaded GIS geographic information and historical inspection data, and multi-source information is fused and overlaid for display, realizing the integrated presentation of on-site equipment status, spatial location and historical data. S4: Based on the real-time video or playback screen of the S3 fusion display, responds to the operator's touch operation or voice command, accurately selects and marks the defective area of ​​the equipment, and automatically associates the defect information, such as defect type, level, tower number and location information, to form a structured defect record. S5: Throughout the entire inspection process from S1 to S4, the video stream, defect labeling data, and flight control commands are uniformly encrypted and securely transmitted in real time based on the national cryptographic SM4 encryption algorithm to ensure the confidentiality, integrity, and tamper resistance of data transmission. S6: After the inspection task is completed, the system automatically summarizes the videos, images, defect annotations, location information and handling suggestions generated throughout the entire process from S1 to S5, and generates a standardized inspection report that conforms to the power industry standards with one click, realizing a closed loop for the entire inspection business process.

[0005] Furthermore, in step S1, visible light and infrared inspection video streams from the UAV are simultaneously acquired through three parallel interfaces: HDMI IN, 3G-SDI IN, and RTSP network. The system automatically identifies and switches between valid signal sources, adaptively analyzes video resolution, encoding format, and frame rate, and completes synchronous alignment and clock calibration of multiple video signals. This provides a unified benchmark for the original video data source, enabling subsequent real-time decoding and spatiotemporal fusion. Furthermore, in step S2, after the video stream enters the system, FFmpeg completes the decapsulation process and calls the VAAPI hardware decoder to decode the video stream in real time; a DMA-BUF+EGL+OpenGL zero-copy rendering pipeline is constructed to directly map the decoded frames from the GPU memory to textures without going through CPU memory copying, thereby achieving low latency and high smoothness display on the hardened terminal.

[0006] Furthermore, in step S3, the defect information includes defect type, level, tower number, and location information.

[0007] Furthermore, in step S3, the video frame timestamp and the UAV's GPS latitude and longitude coordinates are used as a reference to perform spatiotemporal matching with the pre-loaded GIS line and tower geographic information; Akima spline interpolation is used to achieve accurate alignment of real-time video, geographic information, and historical inspection data; real-time images, flight trajectories, tower locations, historical defects, and channel environment information are integrated and overlaid on the same screen, and map linkage, zooming, and precise positioning are supported.

[0008] Furthermore, in step S4, in response to the operator's touchscreen selection, dragging, or voice input commands, the system performs rectangular selection and marking of the defect area on the real-time video screen or playback video; the system automatically associates information such as defect type, defect level, line name, tower number, GPS coordinates, and shooting time to form a structured defect record.

[0009] Furthermore, in step S5, the real-time video stream is encrypted and transmitted using the SM4-CTR mode, and the defect labeling data and flight control commands are encrypted and protected for integrity using the SM4-GCM mode. Combined with SM2 key negotiation, SM3-HMAC verification, two-way digital certificate authentication, and serial number anti-replay mechanism, the end-to-end secure transmission of data throughout the inspection process is achieved, with no plaintext leakage throughout the entire process.

[0010] Furthermore, in step S6, after the inspection task is completed, the system automatically summarizes the entire process data, including video screenshots, defect markings, location descriptions, tower information, and handling suggestions; and exports a standardized inspection report with one click according to the power industry standard template. The report includes a complete defect list, on-site images, location information, and handling suggestions, which can be directly used for defect review, emergency repair dispatch, and file archiving, realizing closed-loop management of inspection business.

[0011] To solve the above-mentioned technical problems, another technical solution provided by the present invention is: a system for real-time image processing and secure transmission for power plant drone inspection, characterized in that, Multi-channel video access module: used to simultaneously acquire visible light and infrared inspection video streams from UAVs through multiple parallel video interfaces, and complete signal identification, switching and clock calibration; Real-time hardware decoding and zero-copy rendering module: used to perform hardware decoding of video streams and achieve low-latency display through the zero-copy rendering pipeline, ensuring real-time performance and smoothness of the picture; Multi-source spatiotemporal data fusion module: used to spatiotemporally align and fuse real-time video with GIS geographic information and historical inspection data for overlay display, so as to achieve integrated presentation of multi-source information; Defect interactive annotation module: used to select and annotate defects in response to touch or voice commands, and automatically associate defect-related information to form a structured record, supporting rapid on-site defect entry; National Cryptographic Security Transmission Module: Used to perform real-time encryption and secure transmission of video streams, annotation data, and control commands based on the national cryptographic SM4 algorithm, ensuring the confidentiality and integrity of data transmission; Standardized report generation module: This module is used to automatically summarize data and generate a standardized inspection report that conforms to power specifications with one click after the inspection task is completed, thus realizing a closed loop in the inspection business.

[0012] Furthermore, the system supports real-time video display, multi-source data fusion, on-site defect labeling, national cryptographic encryption transmission, and automatic report generation, forming a closed-loop operation with data interconnection and time-series linkage between modules. It can independently complete the entire business processing of power drone on-site inspection.

[0013] The beneficial effects of this invention are as follows: 1. This invention optimizes the end-to-end processing architecture of video acquisition, hardware decoding, and zero-copy rendering, abandoning the inefficient processing mode of traditional CPU forwarding and multiple memory copies, significantly compressing data processing latency, and achieving an end-to-end latency of ≤220ms from video access to terminal display. This completely solves the problem of severe screen lag and delay in traditional inspection systems, and fully meets the high real-time operation requirements of power drone field inspection for real-time observation, defect assessment, and equipment control.

[0014] 2. This invention achieves high-precision spatiotemporal unification and integrated overlay visualization of multi-source heterogeneous data, such as real-time inspection videos, GIS geographic information, historical inspection logs, and equipment defect information, through precise spatiotemporal registration and interpolation alignment algorithms. This breaks down the barriers of fragmented and isolated traditional inspection data, allowing operators to intuitively grasp the full-domain information of inspection points, equipment status, historical hazards, and passage environment. It effectively avoids the limitations of single video analysis and significantly improves the accuracy of defect identification and operational decision-making in the field.

[0015] 3. This invention innovatively integrates touch-based selection, drag-and-drop annotation, and voice input into a dual interactive annotation mode, supporting rapid defect annotation in both real-time and playback video scenarios, eliminating the need for post-processing manual data entry and processing. Compared to the traditional manual post-processing annotation method, defect annotation efficiency is improved by over 75%, significantly reducing manual workload and the probability of human error, streamlining post-processing data processing, and enabling real-time collection, archiving, and immediate traceability of defect information.

[0016] 4. This invention adopts the national commercial cryptographic algorithm system throughout, relying on SM4 hierarchical encryption, SM2 secure key negotiation, and SM3 hash verification, combined with two-way certificate authentication and serial number anti-replay mechanism to build an end-to-end comprehensive secure transmission system. It can effectively prevent the leakage, tampering, forgery and replay attack risks of core data such as inspection videos, control commands, and defect ledgers, fully meet the requirements of the power industry network security level protection standard, and adapt to the secure transmission and storage needs of power confidential inspection data.

[0017] 5. This invention integrates the entire business chain from video acquisition, data processing, defect labeling, secure transmission, to report generation. After the inspection task is completed, it can automatically summarize all dimensions of operation data and generate standardized inspection reports with one click based on standard templates in the power industry, eliminating the need for manual statistics, organization, and compilation. Compared with the traditional manual organization and archiving mode, data processing time is reduced by more than 90%, and the overall inspection operation efficiency is improved by more than 29%, realizing closed-loop management of the entire process of "collection-analysis-labeling-archiving-dispatch" for drone inspection operations.

[0018] 6. This invention is compatible with ruggedized hardware operating platforms and can operate stably and continuously in a wide temperature range of -20℃ to 55℃. It also has IP54 dustproof and waterproof protection capabilities, which can effectively resist the effects of harsh working conditions such as wind and sand, rain and snow, high and low temperatures, and complex electromagnetic interference in the field. It is perfectly adapted to complex inspection scenarios such as power lines, mountains, and uninhabited areas in the field, ensuring stable and reliable inspection operations in all weather and all environments.

[0019] To make the above and other objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only two examples of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 Flowchart of the overall implementation of this invention.

[0022] Figure 2 A schematic diagram of the data structure for generating standardized inspection reports with one click. Detailed implementation method: Embodiments of the invention will now be described in detail with reference to the accompanying drawings. While some embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the invention. It should be understood that the drawings and embodiments of the invention are for illustrative purposes only and are not intended to limit the scope of protection of the invention.

[0023] This method utilizes multispectral laser detection to detect contamination on the insulator surface. It combines a physical model with a data-driven model to achieve precise decoupling of salt density, ash density, and moisture content. A closed-loop self-calibration mechanism ensures the stability and accuracy of long-term monitoring. Adaptable to typical polluted environments such as industrial dust, windblown sand and saline-alkali soil, and agricultural dust, this method enables long-term stable online monitoring of transmission line pollution levels. The specific implementation steps are as follows: Example

[0024] like Figure 1-2 As shown, a real-time image processing and secure transmission method for power grid unmanned aerial vehicle (UAV) inspection is disclosed, characterized by the following steps: S1: Synchronously acquire real-time visible light and infrared video streams from UAV inspections through multiple video access interfaces, establish timestamps and spatial coordinate benchmarks for video data, and provide raw data sources for subsequent real-time processing and spatiotemporal fusion. S2: Based on the acquired video stream, the hardware decoder is called to complete real-time decoding, and the decoded image is directly mapped to the ruggedized terminal for low-latency display through the zero-copy rendering pipeline, providing real-time image support for on-site observation and defect identification. S3: Based on the real-time video frame timestamps and UAV positioning information output by S2, the real-time video footage is aligned spatiotemporally with the pre-loaded GIS geographic information and historical inspection data, and multi-source information is fused and overlaid for display, realizing the integrated presentation of on-site equipment status, spatial location and historical data. S4: Based on the real-time video or playback screen of the S3 fusion display, responds to the operator's touch operation or voice command, accurately selects and marks the defective area of ​​the equipment, and automatically associates the defect information, such as defect type, level, tower number and location information, to form a structured defect record. S5: Throughout the entire inspection process from S1 to S4, the video stream, defect labeling data, and flight control commands are uniformly encrypted and securely transmitted in real time based on the national cryptographic SM4 encryption algorithm to ensure the confidentiality, integrity, and tamper resistance of data transmission. S6: After the inspection task is completed, the system automatically summarizes the videos, images, defect annotations, location information and handling suggestions generated throughout the entire process from S1 to S5, and generates a standardized inspection report that conforms to the power industry standards with one click, realizing a closed loop for the entire inspection business process.

[0025] In step S1, three parallel heterogeneous video interfaces—HDMI IN, 3G-SDI IN, and RTSP network—are used to simultaneously acquire dual inspection data streams of visible light high-definition video and infrared thermal imaging video transmitted from the UAV.

[0026] During the access process, the connectivity, bitstream integrity, and transmission stability of each input signal are monitored and evaluated in real time. The system automatically identifies, judges, and switches to the current optimal effective signal source according to preset priorities to ensure that the video link is uninterrupted and the screen does not go black. At the same time, it adaptively parses and matches the video resolution, encoding protocol, encapsulation format, and real-time frame rate to complete the timing synchronization, phase alignment, and global clock calibration between multiple video streams. This constructs a unified spatiotemporal reference for the entire system, thereby providing a reliable, time-consistent, high-quality, frame-free, and low-latency original video data source for subsequent real-time hardware decoding, spatiotemporal fusion of multi-source data, accurate defect labeling, and interactive analysis.

[0027] In step S2, the standardized video stream, after being accessed and calibrated in step S1, enters the system processing link. First, the FFmpeg audio and video processing engine completes the full decapsulation, bitstream parsing, frame segmentation, and format separation of the stream data, accurately extracting the H.264 / H.265 compressed bitstream from the video payload. Then, the compressed bitstream is directly sent to the Intel integrated graphics VAAPI hardware decoder for pure hardware-accelerated real-time decoding, completely avoiding the high utilization rate, high power consumption, and latency fluctuations caused by CPU software decoding, ensuring that the decoding process is efficient and stable.

[0028] Building upon this foundation, the system constructs an integrated zero-copy rendering pipeline of DMA-BUF+EGL+OpenGL. Through the DMA direct memory access mechanism, the raw video frames output by hardware decoding are directly mapped from the GPU memory space to OpenGL sampleable texture objects. The entire process avoids CPU memory transfer, redundant data copying, and format conversion, minimizing data transmission overhead and image processing latency. Ultimately, it achieves low-latency, high-smoothness, tear-free, stutter-free, and screen-flicker-free high-definition real-time video rendering and display on a ruggedized display terminal, providing clear, stable, and reliable image support for subsequent steps such as on-site defect observation, precise bounding box annotation, and real-time control decisions.

[0029] In step S3, the defect information includes defect type, level, tower number, and location information.

[0030] In step S3, a unified spatiotemporal reference benchmark is constructed by combining the timestamps of video frames with the GPS latitude and longitude coordinates transmitted back by the UAV in real time. The dynamic video data collected on-site is associated and bound with the GIS line topology and tower location geographic coordinate information pre-stored locally on the equipment. The Akima spline interpolation algorithm is used to correct temporal misalignment and spatial offset problems, achieving high-precision spatiotemporal alignment of multi-source heterogeneous data such as real-time video, geographic layers, and historical inspection records. The system integrates inspection footage, UAV flight trajectories, power tower distribution points, historical defect records, and information on the surrounding environment of the line, displaying multiple layers of information overlay on a single visual interface. It is also equipped with map linkage switching, multi-level zoom browsing, and precise target location capabilities, comprehensively and intuitively restoring the overall operational status of the inspection area, providing complete spatial data support for defect assessment and operation analysis.

[0031] In step S4, the system can respond in real time to various interactive operation commands such as touch screen selection, drag-and-drop delineation, and voice input. It can operate in the online real-time inspection screen and also retrieve historical playback videos for review and analysis, and complete precise rectangular selection and marking of abnormal defect areas in power equipment.

[0032] After the annotation action takes effect, the system automatically retrieves and matches the corresponding associated data, and synchronously binds core information such as defect category, safety risk level, line name, corresponding tower number, spatial GPS positioning coordinates, and on-site image capture time. It organizes and summarizes all associated parameters, and finally generates a structured defect file record that is formatted in a standardized manner, provides detailed information, allows for full traceability, is easy to retrieve, and is suitable for statistical analysis needs.

[0033] In step S5, the system adopts differentiated national cryptographic encryption strategies to achieve hierarchical security protection based on the transmission characteristics and security requirements of different types of inspection data.

[0034] Specifically, for high-capacity, real-time video streams from UAV inspections, the SM4-CTR streaming encryption mode is used for uninterrupted encrypted transmission. This mode is adapted to the continuous transmission characteristics of video streams and can eliminate the risk of video data leakage in plaintext transmission while ensuring smooth real-time video without stuttering or delay. For critical business and control data such as defect annotation structured data and UAV flight control commands, the SM4-GCM encryption mode with integrity verification capabilities is used to simultaneously complete data encryption, integrity verification, and anti-tampering protection, ensuring the secure and authentic transmission of core business data.

[0035] Meanwhile, the system integrates multiple national cryptographic security protection mechanisms. It completes secure key negotiation between the two communicating parties through the SM2 asymmetric encryption algorithm, realizing secure key interaction and dynamic updates. It relies on the SM3-HMAC hash verification algorithm to perform hash verification on transmitted data, accurately identifying data tampering, loss, and anomalies. It is equipped with a two-way digital certificate authentication mechanism to complete two-way identity verification between devices and terminals, preventing unauthorized devices from accessing the system. It also avoids network attack risks such as data interception and repeated transmission through a serial number anti-replay mechanism.

[0036] By deeply integrating multi-layered security mechanisms, a comprehensive and three-dimensional end-to-end secure transmission system is constructed to achieve secure transmission of all types of data, including drone inspection video streams, defect labeling data, and flight control commands. There is no plaintext data leakage throughout the operation, fully guaranteeing the confidentiality, integrity, authenticity, and uniqueness of power drone inspection data transmission.

[0037] In step S6, after all single inspection tasks are completed, the system automatically collects and integrates various business data generated throughout the operation, such as video screenshots, defect marking logs, point location descriptions, tower basic data, and fault handling suggestions.

[0038] Based on pre-set power industry standard templates, standardized inspection reports with uniform format can be generated and exported with one click. The reports fully include a detailed list of defects, on-site photos, accurate geographical location information, and professional rectification and disposal opinions. The generated reports can be directly used for defect review and approval, on-site emergency repair task assignment, and data archiving and filing, seamlessly connecting various business processes and effectively achieving closed-loop management of the entire inspection operation from data collection and analysis to on-site disposal.

[0039] To solve the above-mentioned technical problems, another technical solution provided by the present invention is: a system for real-time image processing and secure transmission for power plant drone inspection, characterized in that, Multi-channel video access module: It synchronously acquires visible light and infrared dual-channel inspection video streams transmitted by the UAV through multiple parallel heterogeneous video interfaces, judges the signal quality in real time and intelligently switches the effective link, completes the bit stream parsing and timing calibration, and outputs raw video data with unified benchmark. Real-time hardware decoding and zero-copy rendering module: Relying on the hardware decoding architecture to process compressed video streams, and with the zero-copy rendering path to avoid unnecessary data transfer losses, the inspection screen is presented with low latency and high image quality, ensuring real-time and stable operation observation. Multi-source spatiotemporal data fusion module: Using video time sequence and positioning coordinates as a unified reference standard, it completes the spatiotemporal registration of on-site video, GIS geographic layer, and historical inspection archives, and displays multi-layer information overlay with visualization capabilities, while also having map control and precise positioning capabilities. Defect Interactive Annotation Module: Compatible with multiple operation methods such as touch selection, drag-and-drop, and voice input, it can mark equipment defect areas in real-time and playback screens, automatically bind parameters such as line towers, coordinates, time, and defect level, and generate traceable structured defect data; National Cryptographic Security Transmission Module: Based on different data types, it matches differentiated SM4 encryption modes, and works in conjunction with SM2 key negotiation, SM3 verification, identity authentication and anti-replay mechanism to build an end-to-end secure transmission barrier, effectively preventing the risks of data leakage, tampering and unauthorized access; Standardized report generation module: After the inspection is completed, it automatically integrates all business data such as images, defects, locations, and handling suggestions, and automatically generates an inspection report according to the standard template of the power industry, supporting defect review, emergency repair dispatch and data archiving.

[0040] Furthermore, the system enables interconnection and interoperability of data among various functional modules and coordinated operation of runtime sequences. It can complete the entire process of video reception and parsing, multi-dimensional data fusion, defect marking and archiving, encrypted and secure transmission, and inspection report generation in one stop, independently meeting the various operational needs of power drone field inspections.

[0041] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A real-time image processing and secure transmission method for power grid unmanned aerial vehicle (UAV) inspection, characterized in that, Includes the following steps: S1: Synchronously acquire real-time visible light and infrared video streams from UAV inspections through multiple video access interfaces, establish timestamps and spatial coordinate benchmarks for video data, and provide raw data sources for subsequent real-time processing and spatiotemporal fusion. S2: Based on the acquired video stream, the hardware decoder is called to complete real-time decoding, and the decoded image is directly mapped to the ruggedized terminal for low-latency display through the zero-copy rendering pipeline, providing real-time image support for on-site observation and defect identification. S3: Based on the real-time video frame timestamps and UAV positioning information output by S2, the real-time video footage is aligned spatiotemporally with the pre-loaded GIS geographic information and historical inspection data, and multi-source information is fused and overlaid for display, realizing the integrated presentation of on-site equipment status, spatial location and historical data. S4: Based on the real-time video or playback screen of the S3 fusion display, responds to the operator's touch operation or voice command, accurately selects and marks the defective area of ​​the equipment, and automatically associates the defect information, such as defect type, level, tower number and location information, to form a structured defect record. S5: Throughout the entire inspection process from S1 to S4, the video stream, defect labeling data, and flight control commands are uniformly encrypted and securely transmitted in real time based on the national cryptographic SM4 encryption algorithm to ensure the confidentiality, integrity, and tamper resistance of data transmission. S6: After the inspection task is completed, the system automatically summarizes the videos, images, defect annotations, location information and handling suggestions generated throughout the entire process from S1 to S5, and generates a standardized inspection report that conforms to the power industry standards with one click, realizing a closed loop for the entire inspection business process.

2. The real-time image processing and secure transmission method for power plant unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, In step S1, visible light and infrared inspection video streams from the UAV are simultaneously acquired through three parallel interfaces: HDMI IN, 3G-SDI IN, and RTSP network. It automatically identifies and switches valid signal sources, adaptively analyzes video resolution, encoding format and frame rate, and completes the synchronization alignment and clock calibration of multiple video signals, providing a unified benchmark of original video data source for subsequent real-time decoding and spatiotemporal fusion.

3. The real-time image processing and secure transmission method for power plant unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, In step S2, after the video stream enters the system, FFmpeg completes the decapsulation process and calls the VAAPI hardware decoder to decode the video stream in real time; a DMA-BUF+EGL+OpenGL zero-copy rendering pipeline is constructed to directly map the decoded frames from the GPU memory to textures without going through CPU memory copying, thereby achieving low latency and high smoothness display on the hardened terminal.

4. The real-time image processing and secure transmission method for power plant unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, In step S3, the defect information includes defect type, level, tower number, and location information.

5. A real-time image processing and secure transmission method for power plant unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, In step S3, the video frame timestamp and the UAV's GPS latitude and longitude coordinates are used as a reference to perform spatiotemporal matching with the pre-loaded GIS line and tower geographic information; Akima spline interpolation is used to complete the heterogeneous and accurate alignment of real-time video, geographic information, and historical inspection data; real-time images, flight trajectories, tower locations, historical defects, and channel environment information are integrated and overlaid on the same screen, and map linkage, zooming, and precise positioning are supported.

6. A real-time image processing and secure transmission method for power plant unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, In step S4, in response to the operator's touchscreen selection, dragging, or voice input commands, the system performs rectangular selection and marking of the defect area on the real-time video screen or playback video; the system automatically associates information such as defect type, defect level, line name, tower number, GPS coordinates, and shooting time to form a structured defect record.

7. A real-time image processing and secure transmission method for power plant unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, In step S5, the real-time video stream is encrypted and transmitted using the SM4-CTR mode, and the defect labeling data and flight control commands are encrypted and protected for integrity using the SM4-GCM mode. Combined with SM2 key negotiation, SM3-HMAC verification, two-way digital certificate authentication, and serial number anti-replay mechanism, the end-to-end secure transmission of data throughout the inspection process is achieved, with no plaintext leakage throughout the entire process.

8. A real-time image processing and secure transmission method for power plant unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, In step S6, after the inspection task is completed, the system automatically summarizes the entire process data, including video screenshots, defect markings, location descriptions, tower information, and handling suggestions. A standardized inspection report is exported with one click according to the power industry standard template. The report includes a complete defect list, on-site images, location information, and handling suggestions, which can be directly used for defect review, emergency repair dispatch, and file archiving, realizing closed-loop management of inspection business.

9. A system for real-time image processing and secure transmission for power plant unmanned aerial vehicle (UAV) inspection, characterized in that, Multi-channel video access module: used to simultaneously acquire visible light and infrared inspection video streams from UAVs through multiple parallel video interfaces, and complete signal identification, switching and clock calibration; Real-time hardware decoding and zero-copy rendering module: used to perform hardware decoding of video streams and achieve low-latency display through the zero-copy rendering pipeline, ensuring real-time performance and smoothness of the picture; Multi-source spatiotemporal data fusion module: used to spatiotemporally align and fuse real-time video with GIS geographic information and historical inspection data for overlay display, so as to achieve integrated presentation of multi-source information; Defect interactive annotation module: used to select and annotate defects in response to touch or voice commands, and automatically associate defect-related information to form a structured record, supporting rapid on-site defect entry; National Cryptographic Security Transmission Module: Used to perform real-time encryption and secure transmission of video streams, annotation data, and control commands based on the national cryptographic SM4 algorithm, ensuring the confidentiality and integrity of data transmission; Standardized report generation module: This module is used to automatically summarize data and generate a standardized inspection report that conforms to power specifications with one click after the inspection task is completed, thus realizing a closed loop in the inspection business.

10. The on-site real-time image processing and secure transmission system for power plant unmanned aerial vehicle (UAV) inspection according to claim 9, characterized in that, The system supports real-time video display, multi-source data fusion, on-site defect labeling, national cryptographic encryption transmission, and automatic report generation, forming a closed-loop operation with data interconnection and time-series linkage between modules. It can independently complete the full business processing of power drone on-site inspection.

Images