Dynamic dimming method and system for harbor storage yard

Abstract

The invention relates to the technical field of port illumination adjustment, in particular to a port storage yard dynamic dimming method fusing an astronomical algorithm and operation scene recognition, which comprises the following steps: constructing a background illumination model dynamically changing along with time; constructing a storage yard operation situation thermodynamic diagram; constructing a demand mapping rule base, and performing strategy matching on the mapping rule base and the job situation thermodynamic diagram; and constructing a lamp control model, and introducing multi-objective optimization calculation to generate a final control instruction. According to the port storage yard illumination multi-dimensional dynamic dimming method integrating the astronomical algorithm and the operation scene recognition, a cooperative control technology integrating the astronomical algorithm (time dimension) and the AI scene recognition (space dimension) is adopted, only low power is maintained in a storage yard operation area in the daytime, operation illumination is automatically improved to a high-activity area at night, energy conservation and safety are both considered, and the method is suitable for large-scale popularization and application. And a dual smoothing mechanism of smoothing in the morning and evening transition period and real-time dimming flicker prevention is designed, so that sudden brightness change or flicker is avoided, and the visual fatigue of operators is effectively reduced.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology for port lighting, and in particular to a dynamic dimming method and system for port yards. Background Technology

[0002] Port yards are the core areas for container loading, unloading, handling, and storage. Due to the 24-hour operation of port yards, they require sufficiently high levels of illumination (typically requiring an average ground illumination of 20-50 lux or even higher) to ensure that drivers and workers can clearly identify container numbers, locks, and the surrounding environment. At the same time, the illumination in port yards must be uniform to avoid significant alternations between light and dark, which can cause visual fatigue.

[0003] In existing technologies, port yard lighting often employs timed control, zoned constant illumination, or dimming methods based on simple light sensors, as shown in Chinese Patent CN112883548A. This invention belongs to the field of tunnel lighting technology and specifically discloses a dynamic tunnel dimming method, including the following steps: S1: Establishing a dynamic tunnel dimming system; S2: Establishing a dynamic tunnel dimming model within the dynamic tunnel dimming system; S3: Performing dynamic dimming using the dynamic tunnel dimming model and updating the dynamic tunnel dimming model. This invention solves the problems of low intelligence, lack of pre-brightening functionality, and constant lighting duration in existing technologies.

[0004] Existing technical solutions still have many problems in practical applications. Due to the neglect of the impact of changes in ambient natural light on dimming, the difficulty of visual adaptation caused by the difference in brightness between indoors and outdoors, and the lack of a smooth brightness transition design, it is easy to cause sudden brightness changes (such as when a vehicle enters a non-constant area, the brightness jumps directly from a constant value to the calculated value). Long-term passage can easily cause visual fatigue and discomfort. Therefore, based on the above-mentioned technical shortcomings, a dynamic dimming method for port yard lighting is proposed. Summary of the Invention

[0005] Therefore, it is necessary to provide a method and system for dynamic lighting in port yards to address the aforementioned technical problems.

[0006] In a first aspect, this application provides a method for dynamic lighting adjustment in port yards, the method comprising: Obtain the port's geographic coordinates and time data, calculate the solar altitude angle based on the geographic coordinates and time data, and determine the time phase range based on the solar altitude angle; An ambient natural illumination curve is constructed based on the solar altitude angle and time phase range. The ambient natural illumination curve includes time and the basic lighting reference value corresponding to the time. A relationship model between luminaires and lighting areas is constructed based on basic lighting reference values. A multi-objective balance equation is constructed based on basic lighting reference values. The balance equation includes at least the basic illuminance reference value, total energy consumption, illuminance, illuminance uniformity, and dimming range. The equilibrium equation is solved using an iterative greedy algorithm, and the solution results are used to generate dimming commands for the lamps. The dimming commands are then used to control the dimming of the lamps.

[0007] In one embodiment, constructing a multi-objective equilibrium equation based on a baseline lighting value includes: Acquire video footage of the port yard and perform preprocessing based on the video images; Based on the YOLOv8 algorithm model, target recognition is performed on the preprocessed video image and a list of detection boxes is output. The list of detection boxes includes at least the device type, bounding box coordinates, and confidence score. The bounding box list is deduplicated using a nonmaximum suppression algorithm; Predict the position of the target within the detection box based on the Kalman filter algorithm; The predicted target is matched with the detected target based on the Hungarian algorithm, and the matched continuous trajectory is output. Output the target status based on device type, bounding box coordinates, and continuous trajectory; The target state is input into the LSTM time series model and a complex behavior category is output. The complex behavior category includes at least device identity information, behavior type and confidence level. Based on complex behavior categories, predicted target locations, and predicted target trajectories, a heatmap is output using Gaussian kernel density estimation. The brightness of pixels in the heatmap is used to represent the real-time operation intensity at that location. A requirement rule base is constructed based on preset requirements. The requirement rule base includes at least safety illumination requirements, operation accuracy requirements, energy efficiency optimization requirements, and human comfort requirements. The rule-based reasoning matching algorithm matches the real-time situational heat map with the rule base to generate target illuminance value requirements for each activated area. A multi-objective equilibrium equation is constructed based on the target illuminance value requirement and the basic lighting reference value.

[0008] In one embodiment, the heatmap output based on Gaussian kernel density estimation of complex behavior categories, predicted target locations, and predicted target trajectories includes: The Gaussian kernel density estimation expression is as follows: ; Where N is the number of targets identified in the current frame. The weight of the k-th objective is ( , Let K(u, v) be the coordinates of the k-th target on the map, K(·) be the two-dimensional Gaussian kernel function, and K(u, v) = (1 / (2π)). exp(-(u²+v²) / 2), , This is the bandwidth parameter.

[0009] In one embodiment, a rule-based reasoning matching algorithm matches a real-time situational heatmap with a rule base, generating target illuminance values ​​for each activated region. The requirements include: Obtain a scene metadata list, which includes at least device type, device status, device location, device trajectory, and regional identity identifier; When the workload of a grid exceeds a preset activation threshold, the target grid is marked as an active grid. Output the corresponding regional demand list based on each active grid or an active region composed of multiple grids, analyze the scene metadata and filter out the highest priority scene as the dominant scene; Output a list of regional requirements, which includes at least the regional identifier, dominant scene, target illumination, and priority.

[0010] In one embodiment, when a grid or region is simultaneously affected by multiple scenes with different weights, a demand fusion algorithm based on weighted average is used for conflict optimization, as shown in the following formula: ; in, The target illuminance after fusion, Let be the target illumination for the k-th scene. The weight of the k-th scene In one embodiment, when the requirements of two scenarios cannot be easily merged, if the priority of one scenario is higher than the priority of the other scenario, the scenario with lower priority is ignored. A time-domain filter is used for smooth transition processing. The formula for the first-order low-pass filter is: ; in, The target illuminance to be used at the current moment. The original target illuminance calculated at the current moment. The final target illuminance at the previous moment, This is a smoothing factor.

[0011] In one embodiment, the final controlled illuminance value of the luminaire is determined by the maximum value between the basic lighting reference value and the optimized operational requirement illuminance.

[0012] In one embodiment, the model for the relationship between luminaires and illuminated areas based on basic lighting reference values ​​includes: Define the luminaire parameter information, which includes at least the luminaire's identification, luminaire power, luminaire illuminance, luminaire installation coordinates, and adjustable dimming range; Divide the storage yard into grids and set corresponding illuminance requirements; A relationship model between the luminaire and the illuminated area is constructed based on the distance from the luminaire to the center of the area, the angle between the luminaire's light and the area's normal, and the luminaire's illumination angle.

[0013] In one embodiment, solving the equilibrium equation based on an iterative greedy algorithm and generating a lamp dimming command based on the solution includes: Set the dimming level of the lighting fixtures as the decision variable; Construct an objective function based on minimizing total energy consumption; A multi-objective weighted objective function is constructed by considering illuminance satisfaction constraints, illuminance uniformity constraints, and dimming range constraints. The dimming vector of each lamp is optimized based on the objective function and the control command of each lamp is output. The control command includes at least lamp identification information, dimming level, timestamp and command priority. The dimming command is smoothed based on the maximum dimming rate and the ramp function to generate the lamp dimming command.

[0014] Secondly, this application also provides a dynamic dimming system for port yard lighting, the system comprising: The time phase range calculation module is used to obtain the port's geographical coordinates and time data, calculate the solar altitude angle based on the geographical coordinates and time data, and determine the time phase range based on the solar altitude angle; The illumination curve construction module is used to construct an ambient natural illumination curve based on the solar altitude angle and time phase range. The ambient natural illumination curve includes time and the basic lighting reference value corresponding to the time. The equilibrium equation construction module is used to construct a relationship model between luminaires and lighting areas based on basic lighting reference values, and to construct multi-objective equilibrium equations based on basic lighting reference values. The equilibrium equations include at least the basic illuminance reference value, total energy consumption, illuminance, illuminance uniformity, and dimming range. The dimming command output module is used to solve the equilibrium equation based on an iterative greedy algorithm and generate dimming commands for the lamps based on the solution results, and control the dimming of the lamps based on the dimming commands.

[0015] The port yard dynamic lighting method and system provided in this application have the following beneficial effects: 1. The port yard dynamic lighting method and system provided in this application solves the problem of visual adaptation difficulties caused by ignoring changes in natural light by introducing astronomical algorithms to dynamically calculate the basic lighting benchmark; it achieves on-demand lighting allocation by generating a heat map of the operational situation based on video analysis and matching it with a rule base, overcoming the deficiency of traditional lighting methods in lacking scene awareness; it minimizes total energy consumption while ensuring illuminance and uniformity by constructing and solving a multi-objective optimization model; finally, it effectively avoids sudden brightness changes and flickering through a dual smoothing mechanism (dawn / dusk smoothing mapping and command ramp processing), reducing visual fatigue of operators. Experiments show that compared with traditional constant-on or simple zoned timed lighting schemes, this invention can reduce port yard lighting energy consumption by approximately 30%-50% while ensuring operational safety and accuracy.

[0016] 2. Integrating Spatiotemporal Dual-Dimensional Perception: The basic lighting benchmark (temporal dimension) is dynamically calculated through astronomical algorithms, and a heat map of the work situation (spatial dimension) is generated by combining video analysis, so as to achieve adaptive matching between lighting strategies and natural light and work scenarios; 3. Intelligent scene recognition and demand mapping: Based on algorithms such as YOLOv8, Kalman filtering, and LSTM, device tracking and behavior recognition are realized, and refined illuminance requirements are generated through rule matching to improve the accuracy of lighting control and scene adaptability; 4. Multi-objective optimization and energy saving: Construct a multi-objective balance equation with minimizing total energy consumption as the core, while taking into account illuminance satisfaction, uniformity, and dimming range. Solve the equation using an iterative greedy algorithm to significantly reduce energy consumption while ensuring operational safety and visual comfort. 5. Dual smoothing to prevent sudden changes: By using a smoothing mapping function for the transition between dawn and dusk and a dimming command ramp, brightness jumps and flicker are effectively avoided, improving visual comfort and system stability. 6. Strong system scalability: The modular design supports rule base updates and algorithm upgrades, and can be adapted to different port layouts and operation modes, with good engineering applicability; 7. Modular System Architecture: Through the collaboration of six modules—temporal range calculation, ambient lighting curve construction, video perception and situational analysis, rule matching and requirement generation, multi-objective optimization solution, and smooth command issuance—a clear and scalable "perception-decision-execution" closed-loop control system is constructed, improving the efficiency of the method's engineering implementation and the system's reliability. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a dynamic dimming method for port yard lighting in one embodiment; Figure 2 This is a schematic diagram of a dynamic dimming system for port yard lighting in one embodiment; Figure 3This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] This application provides a dynamic dimming method for port yards. This method can be applied to the lighting control system of port yards. The lighting control system stores all the lamps in the port yard, as well as the identity information and brightness information of each lamp. The identity information of the lamp should include the location information of the lamp. The brightness information of the lamp includes not only the maximum power value of the lamp, but also the maximum illuminance and the corresponding dimming range of the lamp.

[0020] The dynamic dimming method for port yard lighting provided in this application can be executed by computing equipment deployed in the port yard control center. This method constitutes a closed-loop control logic of "perception-decision-execution". Specifically: the perception layer obtains a time-varying natural light reference (temporal dimension) through astronomical calculations and obtains a space-varying heatmap of operational intensity (spatial dimension) through video analysis; the decision layer matches and integrates the above spatiotemporal dual-dimensional information with a preset rule base to generate refined target illuminance requirements for each area, and then constructs a multi-objective optimization model with energy saving as the core and balancing illuminance and uniformity, solving for the optimal dimming command for each lamp; the execution layer smooths the command and sends it to the lamps, achieving dynamic and precise adjustment of lighting brightness to adapt to ambient natural light and operational activity intensity.

[0021] The following will take the lighting control system as an example to illustrate... Figure 1 The following is a detailed explanation of a dynamic lighting method for port storage yards, which may include the following steps: Step 101: Obtain the geographical coordinates and time data of the port, calculate the solar altitude angle based on the geographical coordinates and time data, and determine the time phase range based on the solar altitude angle.

[0022] Port geographical coordinates include port longitude. With port dimension Time data includes the current date. Coordinated World Time and the standard time offset of the port's time zone ; The specific steps for calculating the sun's position and determining twilight are as follows: 1. Calculate the Julian Day and the Sun's orbital parameters: The Julian Day (JD) was calculated as a unified astronomical time standard.

[0023] Based on JD, the solar mean ecliptic longitude L, solar mean anomalous angle G, and solar ecliptic longitude are calculated sequentially. obliquity of the ecliptic Solar declination and solar hour angle L and G are used to accurately calculate the solar ecliptic longitude. ; Solar longitude and the obliquity of the ecliptic Used to calculate solar declination Solar declination It is one of the key parameters for calculating the solar altitude angle.

[0024] 2. Calculate the solar altitude angle at key time points: Based on the port's latitude ϕ and the sun's declination With solar hour angle Calculate the solar altitude angle at key times such as sunrise, sunset, and the beginning of dawn and the end of dusk for civilian, maritime, and astronomical purposes. .

[0025] 3. Set the phase range determination logic: The solar altitude angle calculated above The phase range is determined according to the following thresholds: like If the temperature is greater than -0.83°, it is daytime. If -6° < ≤-0.83° is considered civilian day / night time. If -12° < ≤-6°, then it is the twilight or dawn for navigation; If -18° < ≤-12° is considered astronomical twilight; like If the temperature is ≤-18°, it is nighttime.

[0026] The specific calculation formula is as follows: Julian Japan (JD): ; in Y, M, and D represent year, month, and day, respectively.

[0027] Solar longitude (L): ; Sun's horizontal aperimeter angle (G): ; Solar longitude (λ_sun): ; obliquity of the ecliptic ( ): ; Solar declination ( ): ; Solar hour angle (H): ; Local solar time is derived from UTC, Timezone_offset, and longitude ϕ.

[0028] Solar altitude angle (α_sun): ; Step 102: Construct an ambient natural illumination curve based on the solar altitude angle and time phase range. The ambient natural illumination curve includes time and the basic lighting reference value corresponding to the time.

[0029] The specific construction method is as follows: 1) Define the basic lighting pattern: Daytime (α_sun>0°): Baseline value = 0% (off or minimum standby).

[0030] Nighttime (α_sun<-18°): Baseline value = guard level (usually set to 5-10 lx, which meets the minimum safe illuminance requirements for the storage yard in JT / T 557-2023, "Illuminance and Measurement Methods for Port Operation Areas").

[0031] The twilight transition period (-18°≤α_sun≤0°): The baseline value smoothly transitions between the guard level and 0%.

[0032] 2) Construct a smooth mapping function: Standardize the solar altitude angle during the twilight transition period: ; when When = -18°, =0, when When =0°, =1.

[0033] .

[0034] The basic lighting reference value is calculated as follows: Level of protection × (1 - S(x_norm)).

[0035] The curve has the following characteristics: it changes slowly when α_sun is close to -18° and 0°, and changes the fastest near the midpoint (α_sun≈-9°), thus achieving a smooth transition in lighting brightness.

[0036] Step 103: Construct a relationship model between luminaires and lighting areas based on basic lighting reference values, and construct a multi-objective balance equation based on basic lighting reference values. The balance equation includes at least the basic illuminance reference value, total energy consumption, illuminance, illuminance uniformity, and dimming range.

[0037] Step 104: Solve the equilibrium equation based on the iterative greedy algorithm and generate a lamp dimming command based on the solution result, and control the lamp dimming based on the lamp dimming command.

[0038] In this embodiment, by introducing astronomical algorithms to dynamically calculate the basic lighting benchmark, the problem of visual adaptation difficulties caused by ignoring changes in natural light is solved. By generating a heat map of the operational situation based on video analysis and matching it with a rule base, on-demand lighting allocation is achieved, overcoming the lack of scene awareness in traditional lighting methods. By constructing and solving a multi-objective optimization model, the total energy consumption is minimized while ensuring illuminance and uniformity. Finally, through a dual smoothing mechanism (dawn / dusk smoothing mapping and command ramp processing), sudden changes in brightness and flicker are effectively avoided, reducing visual fatigue of operators. Experiments show that compared with traditional constant-on or simple zoned timed lighting schemes, this invention can reduce port yard lighting energy consumption by approximately 30%-50% while ensuring operational safety and accuracy.

[0039] In one embodiment, to address the difficulty of image recognition due to drastic changes in ambient lighting conditions at the storage yard (strong nighttime illumination against a dark background), video stream data from the port storage yard is acquired, and the acquired video images are preprocessed, specifically including: Gaussian filtering is used to reduce image noise; the formula is as follows: ; Where (x,y) are pixel coordinates, σ is the standard deviation used to control the smoothing degree, and I(x,y) is the original image; Histogram equalization is used to enhance contrast and expand the grayscale range of the image. Color correction and white balance processing ensure color consistency under different lighting conditions; All input images are scaled to the fixed size required by the model and then normalized.

[0040] In one embodiment, to perceive the real-time operational status of vehicles, cranes, and other equipment within the port yard, this system employs a pre-trained YOLOv8 object detection model as the basic perception tool. This model can efficiently process video streams and output a list of detection results including equipment type, bounding box coordinates, and confidence scores, providing initial input for subsequent trajectory tracking and behavior analysis. Considering the characteristics of the port yard—open spaces, complex nighttime lighting, and varying target scales—we have adapted the model in the following ways for our application: 1. Data and training: The model is fine-tuned using a dedicated dataset containing a large number of port-specific equipment (such as reach stackers, forklifts, and container trucks) and various lighting conditions (day, night, and dusk) to improve its accuracy and robustness in identifying typical port targets.

[0041] 2. Post-processing: Combining the yard's Geographic Information System (GIS) map, the normalized bounding box coordinates (X_center, Y_center, Width, Height) in the image coordinate system are converted in real time to their positions in the yard's world coordinate system. This is a crucial step in generating subsequent heatmaps for operations. The model's output confidence score and bounding box intersection-over-union (IoU) metrics are primarily used for subsequent non-maximum suppression (NMS) processing to eliminate redundant detection boxes and ensure accurate target localization.

[0042] To eliminate redundant bounding boxes caused by repeated detection of the same target and ensure the accuracy of subsequent tracking and localization, a non-maximum suppression algorithm is used to deduplicate the detection box list. Considering the large structure of large equipment (such as quay cranes and yard cranes) in port yards, which may be segmented in images, the IOU threshold parameter of the NMS algorithm is optimized to more effectively merge adjacent detection boxes belonging to the same physical equipment. This also avoids erroneous merging of different equipment due to close proximity during operations. The resulting clean detection list forms the basis for reliable multi-target tracking and the generation of operational heatmaps.

[0043] The preliminary target information obtained after YOLOv8 detection and coordinate transformation will be input into a multi-target tracking module composed of Kalman filtering and the Hungarian algorithm. By combining a general target detection model with custom tracking and fusion logic, this system constructs a complete situational awareness chain from pixel perception to semantic understanding, which is the foundation for realizing dynamic matching of lighting strategies.

[0044] To achieve continuous, stable, and trajectory-based perception of operating equipment, and to accurately analyze its behavior and predict its future position, this system integrates Kalman filtering and the Hungarian algorithm to construct a robust multi-target tracking module. This module uses Kalman filtering to predict and correct the target motion state, and the Hungarian algorithm to solve the inter-frame target correlation problem. Considering the motion patterns of port yard equipment (such as straight-line travel, frequent starts and stops, and fixed paths) and the high requirements for trajectory smoothness in lighting control, we have made the following key adaptations to the standard algorithm in this application: A prediction and update mechanism based on Kalman filters is used to make optimal estimates of the state (position and velocity) of each tracked target.

[0045] To ensure the accuracy and smoothness of trajectory prediction in the complex environment of the yard, the state transition matrix F and process noise covariance Q of the Kalman filter are specifically configured. For example, for typical linear constant-speed or low-speed moving equipment in the yard, the state vector can be set as [x, y, vx, vy] (position and velocity), and a constant-speed model (CV model) can be used as the basis for the state transition matrix. The process noise covariance Q can be set according to the equipment type. For equipment with relatively high moving speeds, such as container trucks, the Q value can be set higher than that for large equipment with slow moving speeds, such as quay cranes, to reflect the uncertainty of their motion.

[0046] For the matching cost function of the Hungarian algorithm, we define its total cost as the weighted sum of motion association cost and type consistency cost: C_total = β1 C_motion+β2 C_type. The motion association cost C_motion is calculated using Mahalanobis distance; the type consistency cost C_type is defined as follows: if the device type detected in the current frame is the same as the device type predicted by the tracker, then C_type = 0; if they are different, C_type is set to a very large penalty value (e.g., 1000), thereby forcing type matching during the data association stage and effectively reducing the probability of false associations. The weights β1 and β2 can be determined by debugging according to the actual scenario. In this embodiment, they are set to β1 = 0.7 and β2 = 0.3.

[0047] In summary, by configuring Kalman filtering and the Hungarian algorithm specifically for port operations in terms of model parameters and correlation logic, this tracking module can output accurate, stable, and continuous equipment trajectories. This trajectory data is the only reliable input for subsequent behavioral semantic analysis (such as determining loading / unloading, movement, and stationary states) and the generation of operational heatmaps. The "heatmap" generated in this application is not a simple target density map, but a "lighting demand potential field" that integrates equipment type weights, behavioral state weights, and safe operating radius. Ultimately, this ensures that the lighting strategy can accurately match the real-time operational situation, achieving a dynamic balance between safety and energy conservation.

[0048] To stably associate the Kalman filter predictions with the actual detected targets in the current frame, the system employs the Hungarian algorithm for optimal matching. Addressing the association challenges posed by the diverse equipment types and high appearance similarity in port scenarios, we enhanced the standard association cost function: in addition to the Mahalanobis distance cost based on motion consistency, we introduced equipment type matching logic as a strong constraint and utilized the cosine distance of appearance features to assist in re-identification after prolonged occlusion. The specific matching process is as follows: 1. Motion Association Cost: Measured using Mahalanobis distance to evaluate the consistency of motion between the predicted and detected locations, expressed as... ,in, For the j-th detection box, Let i be the predicted position of the i-th tracker. It is the covariance matrix; 2. Appearance Association Cost: This measure of similarity in target appearance features using cosine distance, expressed as... ,in, Let j be the appearance feature vector of the j-th detection box. Let i be the appearance feature library for the i-th tracker.

[0049] By combining the obtained target type, location, trajectory, and behavior, we understand the semantics of the operational scenario. Based on a combination of rules and machine learning, we construct a state recognition logic. The inputs are the target type, location, and motion trajectory. State determination rules are set, including: Determine whether the equipment is in an active operating state based on its type and location; Determine whether the vehicle is in a moving state based on its speed and location; Determine whether the state is stationary based on the idle speed and idle time; Determine whether it is in an idle state based on its location, speed, and idle time; The LSTM time series model is used to analyze short video clips and identify complex behaviors. The formula for the LSTM unit is as follows: Input Gate: ; Forgotten Gate: ; Output gate: ; New candidate value: ; Cell state: ; Hidden state: ; in, As weight, For bias, For the sigmoid function, The input at time t, Let t be the hidden state at time t.

[0050] The target state sequence (including device type, location, speed, and trajectory points) of multiple consecutive frames is input into the LSTM network. Finally, the probability distribution of behavior categories is output through a fully connected layer and a Softmax activation function to identify complex behavior categories such as "loading and unloading operations", "driving", "stationary standby", and "abnormal stop".

[0051] In one embodiment, to transform discrete, dynamic equipment sensing information into a continuous spatial distribution reflecting the overall operational intensity and serving as a direct basis for lighting demand regulation, this system constructs an operational status heatmap generation module based on Gaussian kernel density estimation. This heatmap is not simply a visualization tool, but rather a spatialized demand mapping table for lighting strategies. Its generation process fully considers the physical layout and lighting control characteristics of the port yard, as detailed below: 1) Gridding and Lighting Area Alignment: Based on the actual layout of the yard and the coverage of the lighting fixtures, the plan is divided into regular grids. Each grid corresponds to a minimum independent lighting control unit, ensuring that the granularity of thermal analysis matches the granularity of lighting control.

[0052] 2) Weighted Contribution and Operation Intensity Quantification: The thermal value H(i,j) of each grid is the weighted sum of the contributions of all targets falling within its influence range. Here, the weight w of the target... k It is a key design parameter. It is not a fixed value, but is dynamically adjusted according to the target equipment type (such as quay crane weight > truck weight) and its identified behavior status (such as in operation > moving > stationary), so as to more accurately quantify the actual operation intensity.

[0053] 3) Spatial diffusion based on Gaussian kernel density estimation: The influence of each target is smoothly diffused using a Gaussian kernel function to simulate the impact of operational activities on the lighting requirements of the surrounding area, forming a continuous heat map. Its expression is modified as follows: in, It is a standard two-dimensional Gaussian kernel function.

[0054] The key to compatibility lies in the bandwidth parameter bw x bw y The settings are related to the physical dimensions of typical equipment in the yard, the operational safety radius, and the illumination coverage radius of a single lamp. By adjusting the bandwidth, the natural attenuation range of the impact of operational "hot spots" can be controlled, so that the generated heat map can reflect both the core operational area and the necessary surrounding safety lighting areas, thereby achieving precise pre-allocation of lighting resources.

[0055] Therefore, the heatmap generated by this system integrates multi-dimensional information such as target type, behavioral semantics, spatial location, and safe operating range for each pixel value, and finally outputs a quantified spatial demand matrix H that can directly drive a multi-target lighting optimization model.

[0056] In one embodiment, lighting requirement standards for various work scenarios are predefined, and the requirement dimensions include safety illuminance requirements, work accuracy requirements, energy efficiency optimization requirements, and human comfort requirements, in order to establish a lighting requirement rule base. Safety illuminance requirements are used to limit the minimum illuminance threshold to meet the safe operation of personnel and equipment. Work accuracy requirements are used to limit the target illuminance range corresponding to different work types. Energy efficiency optimization requirements are used to constrain the energy consumption level per unit time under the premise of meeting illuminance requirements. Human comfort requirements are used to limit the rate of change of illuminance and the amplitude of brightness fluctuation in adjacent areas. A rule-based reasoning matching algorithm matches a real-time situational heatmap with a rule base to generate preliminary lighting requirements for each activated region, specifically including: The input scene metadata is List[(device type, status, location coordinates, trajectory, area ID)], and the input situation heatmap is H(i, j), where (i, j) are grid coordinates; Traverse the situational heatmap. For each grid (i, j), if its heat value H(i, j) exceeds the activation threshold T... activate If so, the grid is marked as an active grid; For each active grid or an active region composed of multiple grids, analyze the metadata of all scenes falling within that region, identify the highest-priority scene as the dominant scene for that region, and denote it as the dominant scene. 区域A =argmax s∈S (Priority(s)), where S is the set of all scenarios affecting region A, and Priority(s) is the priority of scenario s retrieved from the rule base; Output a list of regional requirements: List[(Region ID, Dominant Scene, Target Illuminance, Priority)].

[0057] In one embodiment, when a grid or region is simultaneously affected by multiple scenarios with different weights, a demand fusion algorithm based on weighted average is used for conflict optimization. The algorithm formula is as follows: ,in The target illuminance after fusion, Let be the target illumination for the k-th scene. The weight of the k-th scene is proportional to the heat value H(i, j) or the inherent priority of that scene.

[0058] In one embodiment, when the requirements of two scenarios cannot be easily merged, and the priority of one scenario is much higher than that of the other, a priority coverage method is used to resolve the conflict. The conflict resolution strategy is as follows: IF(Scenario A exists AND Scenario B exists) AND(Priority(A) - Priority(B) > ΔP) thresholdTHEN adopts the requirements of scenario A and ignores scenario B, where ΔP threshold Priority difference threshold; A time-domain filter is used for smooth transition processing to prevent frequent lighting flicker caused by brief jitter in target detection or rapid scene switching. The formula for the first-order low-pass filter is: ; in, The target illuminance to be used at the current moment. The original target illuminance calculated at the current moment. The final target illuminance at the previous moment, As a smoothing factor, The smaller the value, the stronger the smoothing effect, but the greater the response delay. Typically, a value of 0.1-0.3 is used, with 0.2 being preferred, to achieve a balance between response speed and smoothness.

[0059] In one embodiment, the final lighting level of a region should simultaneously meet both the environmental safety baseline and the operational activity requirements. The maximum value of the two is taken, and the fusion is performed using the maximum value principle. The formula is as follows: ; in, The final control illuminance value sent to the luminaire. The optimized illumination requirements for the operation. A region-illuminance mapping table is output by fusing the baseline illuminance value related to time t.

[0060] In one embodiment, a physical relationship model between the luminaire and the illuminated area is established, specifically including: The lighting fixture parameters are defined as follows: settings As the unique identifier of the i-th lamp. The maximum power (W) of lamp i is given. Let Lux be the maximum illuminance of luminaire i in a standard plane at 100% luminance. , , () represents the installation coordinates of luminaire i. The dimming range of lamp i is typically [ [,1], where, It has the lowest light ratio; Zone division: Divide the storage yard into M grid zones. j=1,2,...,M, each region Target illumination requirements ; Each luminaire contributes to the illuminance of each area. The illuminance contribution matrix is ​​calculated as follows: C[i,j]=f( , , , ),in, For lighting fixtures to the area Distance from the center The angle between the light ray from the luminaire and the normal to the area. Let C[i,j] be the illumination angle of the lamp, and the simplified model of the illuminance contribution matrix using the inverse square law be C[i,j] = cos( ) / ( +h²), where h is the installation height, cos( ) is a cosine law factor.

[0061] In one embodiment, based on the regional illuminance requirements obtained from the preceding steps (i.e., the target illuminance in the regional requirement list) and the illuminance contribution model between luminaires and regions, the lighting control problem is constructed as a multi-objective optimization problem. The goal of solving this problem is to find an optimal set of luminaire dimming levels that, while satisfying the illuminance requirements and uniformity requirements of all work areas, minimize the total system energy consumption. The specific construction is as follows: Set the decision variable to d. i d i ∈[D min ,1], where d i The dimming level (D) of luminaire i min (This refers to the minimum dimming ratio allowed for the luminaire). With minimizing the total system energy consumption as the primary objective, its basic objective function is: ; in This represents the maximum power of lamp i at 100% brightness.

[0062] The constraints include: 1. Illuminance constraint: The actual illuminance of each area j must not be lower than its required illuminance, i.e. ; in The list of regional requirements from the rule matching output reflects the safety and accuracy requirements of the region based on the real-time operational situation.

[0063] 2. Illuminance Uniformity Constraint: To avoid excessively drastic alternation between light and dark areas in adjacent regions, which could lead to visual fatigue, the ratio of the actual illuminance to the required illuminance in each region j must not exceed the upper limit U. max ,Right now U maxThe settings are based on port operation visual ergonomic standards to ensure a smooth transition in illuminance distribution.

[0064] 3. Dimming Range Constraints: The dimming level of each lamp must be within its adjustable range, i.e. D min ≤d i ≤1,∀i∈{1,...,N} D min The lifespan of a lamp is determined by its physical properties, and frequent switching or excessively low brightness should be avoided to prevent it from being affected.

[0065] To simultaneously optimize energy consumption and illuminance satisfaction, the total penalty for insufficient illuminance is taken as a secondary objective. A comprehensive objective function is constructed and solved by weighting the total energy consumption by λ and summing the weights: Here, λ is a positive weighting coefficient used to balance energy saving and illuminance compliance. The value of λ can be determined through experience or simulation. For example, it can be set to a larger value (e.g., 10) during peak operating periods to prioritize illuminance, and a smaller value (e.g., 1) during off-peak operating periods to prioritize energy saving.

[0066] In one embodiment, for real-time control of large-scale lighting fixtures, an iterative greedy algorithm is used to solve a multi-objective optimization problem to obtain the optimal lighting fixture dimming scheme, including: Initialize all lights to minimum brightness. = ,∀i; while (there exists a region j that satisfies) < ); Find the region with the largest illuminance gap j star Find the lighting fixtures that contribute the most to the area. star Increase by step size =min(1, +Δd); This iteratively improves the brightness of luminaires in areas with insufficient illumination, converting the mathematical solution into actual control commands, specifically: Input: The dimming vector d obtained from the optimized solution d=[ , ,..., ]; Output: Control commands for each light fixture Command format: {Lamp ID, Dimming level, Timestamp, Command priority}; A ramp function is used for smooth transition processing to output the lighting control command set, represented as follows: (t)= +sign(Δd)˙min(|Δd|,R max ˙Δt); Among them, R max Δt represents the maximum dimming rate, and Δt represents the control cycle time.

[0067] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0068] Based on the same inventive concept, such as Figure 2 As shown in the illustration, this application also provides a dynamic dimming system for port yard lighting, the system comprising: The time phase range calculation module is used to obtain the port's geographical coordinates and time data, calculate the solar altitude angle based on the geographical coordinates and time data, and determine the time phase range based on the solar altitude angle; The illumination curve construction module is used to construct an ambient natural illumination curve based on the solar altitude angle and time phase range. The ambient natural illumination curve includes time and the basic lighting reference value corresponding to the time. The equilibrium equation construction module is used to construct a relationship model between luminaires and lighting areas based on basic lighting reference values, and to construct multi-objective equilibrium equations based on basic lighting reference values. The equilibrium equations include at least the basic illuminance reference value, total energy consumption, illuminance, illuminance uniformity, and dimming range. The dimming command output module is used to solve the equilibrium equation based on an iterative greedy algorithm and generate dimming commands for the lamps based on the solution results, and control the dimming of the lamps based on the dimming commands.

[0069] In one embodiment, a computer device is provided, the internal structure of which can be shown as follows: Figure 3As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for data exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a dynamic lighting method for a port yard.

[0070] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0071] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0072] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0073] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0074] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The processors involved in the embodiments provided in this application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited thereto.

[0075] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0076] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for dynamic lighting adjustment in port storage yards, characterized in that, The method includes: Obtain the port's geographic coordinates and time data, calculate the solar altitude angle based on the geographic coordinates and time data, and determine the time phase range based on the solar altitude angle; An ambient natural illumination curve is constructed based on the solar altitude angle and time phase range. The ambient natural illumination curve includes time and the basic lighting reference value corresponding to the time. A relationship model between luminaires and lighting areas is constructed based on basic lighting reference values. A multi-objective balance equation is constructed based on basic lighting reference values. The balance equation includes at least the basic illuminance reference value, total energy consumption, illuminance, illuminance uniformity, and dimming range. The equilibrium equation is solved using an iterative greedy algorithm, and the solution results are used to generate dimming commands for the lamps. The dimming commands are then used to control the dimming of the lamps.

2. The method according to claim 1, characterized in that, The construction of the multi-objective equilibrium equation based on the basic lighting reference value includes: Acquire video footage of the port yard and perform preprocessing based on the video images; Based on the YOLOv8 algorithm model, target recognition is performed on the preprocessed video image and a list of detection boxes is output. The list of detection boxes includes at least the device type, bounding box coordinates, and confidence score. The bounding box list is deduplicated using a nonmaximum suppression algorithm; Predict the position of the target within the detection box based on the Kalman filter algorithm; The predicted target is matched with the detected target based on the Hungarian algorithm, and the matched continuous trajectory is output. Output the target status based on device type, bounding box coordinates, and continuous trajectory; The target state is input into the LSTM time series model and a complex behavior category is output. The complex behavior category includes at least device identity information, behavior type and confidence level. Based on complex behavior categories, predicted target locations, and predicted target trajectories, a heatmap is output using Gaussian kernel density estimation. The brightness of pixels in the heatmap is used to represent the real-time operation intensity at that location. A requirement rule base is constructed based on preset requirements. The requirement rule base includes at least safety illuminance requirements, operation accuracy requirements, energy efficiency optimization requirements, and human comfort requirements. The safety illuminance requirements are used to limit the minimum illuminance threshold for safe operation of personnel and equipment. The operation accuracy requirements are used to limit the target illuminance range corresponding to different operation types. The energy efficiency optimization requirements are used to constrain the energy consumption level per unit time under the premise of meeting the illuminance requirements. The human comfort requirements are used to limit the rate of change of illuminance and the amplitude of brightness fluctuation in adjacent areas. The rule-based reasoning matching algorithm matches the real-time situational heat map with the rule base to generate target illuminance value requirements for each activated area. A multi-objective equilibrium equation is constructed based on the target illuminance value requirement and the basic lighting reference value.

3. The method according to claim 2, characterized in that, The heatmap output based on complex behavior categories, predicted target location, and predicted target trajectory using Gaussian kernel density estimation includes: The Gaussian kernel density estimation expression is as follows: ； in, , The row and column indices are used to represent the grid in the situational heatmap, where N is the number of targets identified in the current frame. The weight corresponding to the k-th target is ( , Let K(u, v) be the coordinates of the k-th target on the map, K(·) be the two-dimensional Gaussian kernel function, and K(u, v) = (1 / (2π)). exp(-(u²+v²) / 2), , Here, u and v represent the spatial offsets in the Gaussian kernel function, which are bandwidth parameters.

4. The method according to claim 2, characterized in that, The rule-based reasoning matching algorithm matches the real-time situational heatmap with the rule base, generating target illuminance values ​​for each activated region, including: Obtain a scene metadata list, which includes at least device type, device status, device location, device trajectory, and regional identity identifier; When the workload of a grid exceeds a preset activation threshold, the target grid is marked as an active grid. Output the corresponding regional demand list based on each active grid or an active region composed of multiple grids, analyze the scene metadata and filter out the highest priority scene as the dominant scene; Output a list of regional requirements, which includes at least the regional identity, dominant scene, target illumination, and priority.

5. The method according to claim 4, characterized in that, The method further includes: When one grid or region is simultaneously affected by multiple scenarios with different weights, a demand fusion algorithm based on weighted average is used for conflict optimization, as shown in the following formula: ； in, To affect the number of scenes in the current grid or region, The target illuminance after fusion, Let be the target illumination for the k-th scene. This represents the weight corresponding to the k-th scene.

6. The method according to claim 5, characterized in that, The method further includes: When the requirements of two scenarios cannot be easily merged, if the priority of one scenario is higher than that of the other, the scenario with the lower priority is ignored. Time-domain filtering is used for smooth transition processing. The formula for a first-order low-pass filter is: ； in, The target illuminance to be used at the current moment. The original target illuminance calculated at the current moment. The final target illuminance at the previous moment, This is a smoothing factor.

7. The method according to claim 6, characterized in that, The method further includes: The final controlled illuminance value of the luminaire is determined by the maximum value between the basic lighting reference value and the optimized operational requirement illuminance.

8. The method according to claim 1, characterized in that, The model for the relationship between luminaires and lighting areas based on basic lighting benchmark values ​​includes: Define the luminaire parameter information, which includes at least the luminaire's identification, luminaire power, luminaire illuminance, luminaire installation coordinates, and adjustable dimming range; Divide the storage yard into grids and set corresponding illuminance requirements; A relationship model between the luminaire and the illuminated area is constructed based on the distance from the luminaire to the center of the area, the angle between the luminaire's light and the area's normal, and the luminaire's illumination angle.

9. The method according to claim 8, characterized in that, The step of solving the equilibrium equation based on an iterative greedy algorithm and generating a lamp dimming command based on the solution result includes: Set the dimming level of the lighting fixtures as the decision variable; Construct an objective function based on minimizing total energy consumption; A multi-objective weighted objective function is constructed by considering illuminance satisfaction constraints, illuminance uniformity constraints, and dimming range constraints. The dimming vector of each lamp is optimized based on the objective function and the control command of each lamp is output. The control command includes at least lamp identification information, dimming level, timestamp and command priority. The dimming command is smoothed based on the maximum dimming rate and the ramp function to generate the lamp dimming command.

10. A dynamic dimming system for port yard lighting, characterized in that, The system includes: The time phase range calculation module is used to obtain the port's geographical coordinates and time data, calculate the solar altitude angle based on the geographical coordinates and time data, and determine the time phase range based on the solar altitude angle; The illumination curve construction module is used to construct an ambient natural illumination curve based on the solar altitude angle and time phase range. The ambient natural illumination curve includes time and the basic lighting reference value corresponding to the time. The equilibrium equation construction module is used to construct a relationship model between luminaires and lighting areas based on basic lighting reference values, and to construct multi-objective equilibrium equations based on basic lighting reference values. The equilibrium equations include at least the basic illuminance reference value, total energy consumption, illuminance, illuminance uniformity, and dimming range. The dimming command output module is used to solve the equilibrium equation based on an iterative greedy algorithm and generate dimming commands for the lamps based on the solution results, and control the dimming of the lamps based on the dimming commands.

Images