Visual monitoring system for hydropower stations in harsh environments and method thereof

Abstract

The application discloses a kind of visual monitoring systems for hydroelectric station under harsh environment and method thereof, including image acquisition unit, environment perception and preprocessing unit, adaptive image enhancement unit, result fusion and output unit and safety monitoring and alarm unit, from system level, complete visual safety monitoring system including image acquisition unit, environment perception and preprocessing unit, adaptive image enhancement unit and safety monitoring unit are constructed, interface is clear between each functional module, structure is clear;By introducing the recursive image enhancement mechanism based on curve modeling, the low-illumination monitoring image is gradually enhanced for several times, the image brightness is gradually improved in a smooth way, and the noise amplification and detail distortion problems commonly existing in the traditional one-time brightness stretching method are avoided, which significantly improves the visibility and stability of the monitoring image under harsh environment.

Description

Technical Field

[0001] This invention relates to the field of hydropower station monitoring technology, specifically to a visual monitoring system and method for hydropower stations in harsh environments. Background Technology

[0002] Cascade hydropower stations, as important clean energy infrastructure, are typically distributed longitudinally along rivers, characterized by long operating cycles, wide monitoring ranges, and complex operating environments. To ensure the safe and stable operation of hydropower stations, long-term and continuous safety monitoring of the powerhouse equipment, water conveyance tunnels, dam structures, and surrounding environment is necessary. Existing hydropower station monitoring systems suffer from the following problems: low illumination and non-uniform lighting; high humidity; water mist and reflection interference; and complex backgrounds superimposed on equipment structures. Hydropower station powerhouses, underground tunnels, and parts of the dam structure are often in enclosed or semi-enclosed environments, with limited lighting conditions at night and during non-working hours. This results in generally low brightness in monitoring images and significant non-uniform lighting, leading to insufficient image contrast and loss of detail. Furthermore, hydropower stations operate in high humidity environments, with significant water vapor, mist, and equipment surface reflections, easily introducing noise, blurring, and locally bright areas into the monitoring images, further exacerbating image degradation. The complex internal equipment structures and dense pipelines of hydropower stations cause background textures to overlap with equipment features. Under low light conditions, traditional monitoring images struggle to accurately distinguish key structures from background information, affecting the reliability of safety monitoring. Existing technologies address lighting issues by adding artificial lighting or infrared equipment, and image enhancement methods based on fixed parameters improve image visibility. However, these methods suffer from several problems: a lack of dedicated visual enhancement solutions for low-light, non-uniform lighting, and high-humidity environments in hydropower stations; fixed enhancement intensity and processes in existing enhancement methods, making it difficult to adapt to dynamic changes in different areas and time periods; and a disconnect between image enhancement and safety monitoring tasks, resulting in a lack of integrated engineering system design that affects overall monitoring effectiveness and stability. Summary of the Invention

[0003] The purpose of this invention is to provide a visual monitoring system and method for hydropower stations in harsh environments, in order to solve the problems mentioned in the background art.

[0004] By adopting the above technical solutions, the visibility and stability of monitoring images in harsh environments are significantly improved; the enhancement results are well adaptable to complex lighting environments; the dependence of the enhancement process on human experience and fixed parameters is reduced; the enhancement results are more suitable for security monitoring and engineering application needs; the system structure is clear and easy to deploy and integrate.

[0005] Compared with existing technologies, the beneficial effects of this invention are as follows: Starting from the system level, a complete visual safety monitoring system is constructed, including an image acquisition unit, an environmental perception and preprocessing unit, an adaptive image enhancement unit, and a safety monitoring unit. The interfaces between each functional module are clear, and the structure is well-defined. By introducing a recursive image enhancement mechanism based on curve modeling, low-light monitoring images are enhanced multiple times in a progressive manner, so that the image brightness is gradually increased in a smooth way. This avoids the noise amplification and detail distortion problems that are common in traditional one-time brightness stretching methods, and significantly improves the visibility and stability of monitoring images in harsh environments. Compared with the prior art, the beneficial effects of the present invention are: the introduction of a multi-head adaptive curve enhancement mechanism dynamically adjusts the weight ratio of each enhancement head according to the illumination characteristics of different regions in the same image, thereby ensuring effective enhancement of dark areas while avoiding over-processing of bright areas; the generated enhanced image not only has a significant improvement in subjective visual effect, but also performs stably in terms of structural information preservation, edge sharpness and noise suppression, which is more in line with the actual image quality requirements of security monitoring tasks. Compared with the prior art, the beneficial effects of the present invention are: the introduction of an adaptive enhanced step size control mechanism can automatically adapt to different hydropower station scenarios without manual intervention, effectively reducing system operation and maintenance costs and improving long-term operational stability. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of the overall structure of the safety monitoring system of the present invention; Figure 2 This is a flowchart illustrating the image acquisition unit of the present invention; Figure 3 This is a schematic diagram of the structure and flow of the environmental perception and preprocessing unit of the present invention; Figure 4 This is a schematic diagram of the structure of the multi-head adaptive curve enhancement module of the present invention. Detailed Implementation

[0007] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0008] Please see Figure 1-4This invention provides a technical solution: a visual monitoring system for hydropower stations in harsh environments, comprising an image acquisition unit, an environmental perception and preprocessing unit, an adaptive image enhancement unit, a result fusion and output unit, and a safety monitoring and alarm unit. The image acquisition unit is used to acquire real-time or periodic images of the operating status of the monitoring area of ​​the cascade reservoir hydropower station, serving as the data input for the entire visual safety monitoring system. The environmental perception and preprocessing unit performs basic processing on the raw image data transmitted by the image acquisition unit and extracts feature information representing the current monitoring environment's lighting conditions. The adaptive image enhancement unit is the core functional module of this invention's visual safety monitoring system, used to adaptively enhance low-quality monitoring images to improve image visibility and structural information integrity. The result fusion and output unit performs unified processing on the multiple enhancement results output by the adaptive image enhancement unit to generate the final enhanced image. The safety monitoring and alarm unit performs safety analysis and anomaly detection based on the enhanced monitoring image, serving as the application output for this invention's visual safety monitoring system. The image acquisition unit may include, but is not limited to, visible light cameras, low-light cameras, or video acquisition devices compatible with existing monitoring systems. These devices can be deployed in key monitoring areas such as underground powerhouses, water conveyance tunnels, dam surfaces, critical equipment areas, and access points of hydropower stations. The image acquisition unit supports multi-channel parallel acquisition, enabling it to simultaneously acquire image data from multiple monitoring areas and transmit the acquired raw image data to the backend processing module via wired or wireless networks. The image acquisition unit itself does not perform complex image quality processing but rather ensures the integrity and continuity of the image data, providing raw input for subsequent environmental perception and enhancement processing. The environmental perception and preprocessing unit can perform preprocessing operations on the input image, including but not limited to size normalization, color space conversion, basic noise suppression, and pixel value range standardization, to improve the processing stability of the subsequent adaptive enhancement module. It analyzes the overall brightness distribution, contrast features, and illumination uniformity of the image, extracting global feature vectors that reflect the current monitoring environment's illumination status. The extracted environmental feature information serves as an important input for the subsequent adaptive image enhancement unit, guiding the dynamic adjustment of enhancement strategies and intensity. By setting up the environmental perception and preprocessing unit, this invention enables the system to possess basic perception capabilities for different monitoring scenarios, laying the foundation for stable enhancement processing in the complex and dynamically changing operating environment of hydropower stations. The adaptive image enhancement unit includes a curve-modeling-based recursive enhancement module, a multi-head adaptive curve enhancement module, and an adaptive enhancement step size control module. The curve-modeling-based recursive enhancement module constructs a pixel-level brightness mapping curve and performs multiple recursive enhancements on the input image, gradually bringing the image brightness closer to the target exposure level, thus performing progressive brightness enhancement processing on low-light monitoring images. The multi-head adaptive curve enhancement module constructs multiple enhancement strategies in parallel, performing multi-path enhancement processing on the same input image to address the non-uniform illumination problem commonly found in hydropower station monitoring images. The adaptive enhancement step size control module adaptively determines the step size required for the enhancement process based on the overall illumination characteristics of the input image, dynamically adjusting the number of enhancement iterations in the recursive enhancement process. The curve-modeling-based recursive enhancement module can… It effectively avoids noise amplification and detail loss caused by sudden brightness changes, making it particularly suitable for monitoring scenarios with extremely poor lighting, such as underground factories and water conveyance tunnels. The multi-head adaptive curve enhancement module adaptively adjusts the weight ratio of each enhancement head based on the lighting feature information extracted by the environmental perception and preprocessing unit, and fuses the multi-channel enhancement results. Through the multi-head adaptive curve enhancement mechanism, the system can take into account the enhancement needs of different areas in the same image, improving the overall enhancement effect in complex lighting environments. In monitoring scenarios with extremely poor lighting conditions, the adaptive enhancement step size control module can automatically increase the number of enhancement iterations to ensure full recovery of details in dark areas; in scenarios with relatively good lighting conditions or with reflection interference, it automatically reduces the number of enhancement iterations to avoid the negative effects of over-enhancement. The result fusion and output unit performs weighted fusion of each enhancement result based on the weight information calculated by the multi-head adaptive curve enhancement module, ensuring that the final output image achieves a balance in terms of brightness, contrast and structural information; the result fusion and output unit is used to uniformly process the multi-channel enhancement results output by the adaptive image enhancement unit to generate the final enhanced image. The safety monitoring and alarm unit can be configured with different safety monitoring functions according to actual needs, including but not limited to equipment status identification, structural anomaly detection, water accumulation or leakage identification, and personnel behavior monitoring. When an anomaly is detected, the safety monitoring and alarm unit can issue alarm information through audible and visual alarms, information push, or linkage with the hydropower station safety management platform. Since the enhanced image is significantly better than the original monitoring image in terms of visibility and stability, the safety monitoring and alarm unit can maintain high detection reliability in harsh environments, effectively reducing the risk of false alarms and missed alarms.

[0009] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0010] Please see Figure 1-4 This invention provides a technical solution: a method for a visual monitoring system for hydropower stations in harsh environments, comprising the following steps: S1. Monitoring Image Acquisition: Raw monitoring images of the hydropower station's operation are acquired through image acquisition devices deployed in the powerhouse, underground caverns, dam surface, and key equipment areas of the cascade reservoir hydropower station. The image acquisition devices can be visible light cameras or video acquisition equipment compatible with existing monitoring systems. S2. Image preprocessing and environmental feature extraction: Preprocessing operations are performed on the acquired raw monitoring images, including size normalization, color space conversion, and basic noise suppression; at the same time, environmental features are extracted from the preprocessed images to obtain global feature information that characterizes the current monitoring environment lighting conditions. S3. Recursive Image Enhancement Based on Curve Modeling: This method employs curve modeling to progressively enhance low-light surveillance images. The enhancement process uses a recursive approach.

[0011] Used to describe the brightness mapping relationship of the input image during a single enhancement process.

[0012] This is used to represent the update form in which the image brightness gradually approaches the target exposure level during multiple recursive enhancement processes; S4. Multi-head adaptive curve enhancement processing: Introduces a multi-head adaptive curve enhancement mechanism to execute multiple enhancement strategies in parallel on the input image;

[0013] Global average pooling is performed on shared features to extract the overall illumination distribution features of the image;

[0014] Calculate the original weight score corresponding to each enhancement head;

[0015] The weight scores are normalized to obtain the weight coefficients of each enhancement head in the final output; finally, the formula is used... The outputs of different enhancement heads are weighted and fused to generate an intermediate enhancement result; S5. Adaptive enhancement step size control: An adaptive enhancement step size control mechanism is introduced to dynamically adjust the number of enhancement iterations based on the overall illumination characteristics of the current image. Describe the traditional enhancement form under fixed iteration conditions;

[0016] Used to extract the global illumination description vector of an image; Map global features to continuous scalars representing enhancement strength; Used to map continuous prediction results to discrete enhancement step sizes; The final number of enhancement iterations is determined by probability sampling. S6. Enhanced output and safety monitoring analysis: The final enhanced image is used as input for safety monitoring analysis to identify the status of key equipment and operating environment of the hydropower station, detect anomalies or conduct safety assessments, and output corresponding monitoring results or early warning information.

Claims

1. A visual monitoring system for hydropower stations in harsh environments, characterized in that: The system comprises an image acquisition unit, an environmental perception and preprocessing unit, an adaptive image enhancement unit, a result fusion and output unit, and a safety monitoring and alarm unit. The image acquisition unit is used to acquire real-time or periodic images of the operational status of the monitoring area of ​​the cascade reservoir hydropower station, serving as the data input for the entire visual safety monitoring system. The environmental perception and preprocessing unit performs basic processing on the raw image data transmitted by the image acquisition unit and extracts feature information representing the current monitoring environment's lighting conditions. The adaptive image enhancement unit is the core functional module of the visual safety monitoring system, used to adaptively enhance low-quality monitoring images to improve image visibility and structural information integrity. The result fusion and output unit performs unified processing on the multiple enhancement results output by the adaptive image enhancement unit to generate the final enhanced image. The safety monitoring and alarm unit performs safety analysis and anomaly detection based on the enhanced monitoring image, serving as the application output for the visual safety monitoring system.

2. The visual monitoring system for hydropower stations in harsh environments according to claim 1, characterized in that: The image acquisition unit may include, but is not limited to, visible light cameras, low-light cameras, or video acquisition devices compatible with existing monitoring systems. The devices may be deployed in key monitoring areas such as underground powerhouses, water conveyance tunnels, dam surfaces, critical equipment areas, and passage entrances and exits of hydropower stations. The image acquisition unit supports multi-channel parallel acquisition, which can simultaneously acquire image data from multiple monitoring areas and transmit the acquired raw image data to the back-end processing module via wired or wireless networks.

3. The visual monitoring system for hydropower stations in harsh environments according to claim 1, characterized in that: The environmental perception and preprocessing unit can perform preprocessing operations on the input image, including but not limited to size normalization, color space conversion, basic noise suppression, and pixel value range standardization, to improve the processing stability of the subsequent adaptive enhancement module; it can also analyze the overall brightness distribution, contrast features, and illumination uniformity of the image to extract a global feature vector that reflects the current ambient illumination status.

4. The visual monitoring system for hydropower stations in harsh environments according to claim 1, characterized in that: The adaptive image enhancement unit includes a curve-modeling-based recursive enhancement module, a multi-head adaptive curve enhancement module, and an adaptive enhancement step size control module. The curve-modeling-based recursive enhancement module constructs a pixel-level brightness mapping curve and performs multiple recursive enhancements on the input image, so that the image brightness gradually approaches the target exposure level for performing progressive brightness enhancement processing on low-light monitoring images. The multi-head adaptive curve enhancement module solves the common problem of non-uniform lighting in hydropower station monitoring images by constructing multiple enhancement strategies in parallel and performing multi-path enhancement processing on the same input image. The adaptive enhancement step size control module adaptively determines the step size required for the enhancement process based on the overall illumination characteristics of the input image, thereby dynamically adjusting the number of enhancement iterations in the recursive enhancement process.

5. A visual monitoring system for hydropower stations in harsh environments according to claim 1, characterized in that: The result fusion and output unit performs weighted fusion of each enhancement result based on the weight information calculated by the multi-head adaptive curve enhancement module, so as to ensure that the final output image achieves a balanced state in terms of brightness, contrast and structural information.

6. A visual monitoring system for hydropower stations in harsh environments according to claim 1, characterized in that: The safety monitoring and alarm unit can be configured with different safety monitoring functions according to actual needs, including but not limited to equipment status identification, structural anomaly detection, water accumulation or leakage identification, and personnel behavior monitoring. When an abnormal situation is detected, the safety monitoring and alarm unit can issue alarm information through audible and visual alarms, information push, or linkage with the hydropower station safety management platform.

7. A method for a visual monitoring system for hydropower stations in harsh environments, as described in claims 1-6, characterized in that: Includes the following steps: S1. Monitoring Image Acquisition: Raw monitoring images of the hydropower station's operation are acquired using image acquisition devices deployed in the powerhouse, underground chambers, dam surface, and key equipment areas of the cascade reservoir hydropower station. The image acquisition devices can be visible light cameras or video acquisition equipment compatible with existing monitoring systems. S2. Image preprocessing and environmental feature extraction: Preprocessing operations are performed on the acquired raw monitoring images, including size normalization, color space conversion, and basic noise suppression; at the same time, environmental features are extracted from the preprocessed images to obtain global feature information that characterizes the current monitoring environment lighting conditions. S3. Recursive image enhancement based on curve modeling: This method uses curve modeling to progressively enhance low-light surveillance images. The enhancement process is recursive: ； Used to describe the brightness mapping relationship of the input image during a single enhancement process. ； This is used to represent the update form in which the image brightness gradually approaches the target exposure level during multiple recursive enhancement processes; S4. Multi-head adaptive curve enhancement processing: Introduces a multi-head adaptive curve enhancement mechanism to execute multiple enhancement strategies in parallel on the input image; ； Global average pooling is performed on shared features to extract the overall illumination distribution features of the image; ； Calculate the original weight score corresponding to each enhancement head; ； The weight scores are normalized to obtain the weight coefficients of each enhancement head in the final output; finally, the formula is used... The outputs of different enhancement heads are weighted and fused to generate an intermediate enhancement result; S5. Adaptive enhancement step size control: An adaptive enhancement step size control mechanism is introduced to dynamically adjust the number of enhancement iterations based on the overall illumination characteristics of the current image. Describe the traditional enhancement form under fixed iteration conditions; ； Used to extract the global illumination description vector of an image; Map global features to continuous scalars representing enhancement strength; Used to map continuous prediction results to discrete enhancement step sizes; The final number of enhancement iterations is determined by probability sampling. S6. Enhanced output and safety monitoring analysis: The final enhanced image is used as input for safety monitoring analysis to identify the status of key equipment and operating environment of the hydropower station, detect anomalies or conduct safety assessments, and output corresponding monitoring results or early warning information.

Images