Classroom lighting partition intelligent adjustment system and method based on personnel position and number statistics

Abstract

This invention discloses a classroom lighting zone intelligent adjustment system and method based on personnel location and headcount statistics. The system includes an image acquisition layer, an edge recognition layer, a region mapping and statistics layer, a lighting rule decision layer, a lighting control execution layer, and a data management and interaction layer. The image acquisition layer is responsible for acquiring images or video frames of the classroom. The edge recognition layer detects student faces or heads at the edges and outputs the target location. The region mapping and statistics layer maps the detected targets to pre-defined lighting areas and then counts the number of people in each area. The lighting rule decision layer determines the target brightness based on the brightness rules for the number of people, generates the target brightness, the lighting control execution layer drives the lights in the corresponding areas, and the data management and interaction layer records the detection results and lighting operation logs. This invention can address problems such as all-on / all-off classroom lighting, lack of personnel distribution perception, and coarse-grained zone brightness control, achieving on-demand dimming and zoned energy-saving control.

Description

Technical Field

[0001] This invention belongs to the fields of smart classroom, intelligent lighting, computer vision, edge computing and Internet of Things control technology, and specifically relates to a classroom lighting zone intelligent adjustment system and method based on personnel location and number statistics. Background Technology

[0002] With the ongoing development of smart campuses, smart stadiums, and energy-saving renovations of public buildings, classroom lighting systems are gradually evolving from traditional manual switches towards automated and intelligent control. Currently, existing classrooms typically use manual switches, timer controls, infrared motion sensors, or simple light sensor linkages to control the lights. While this reduces the burden of manual management to some extent, it remains difficult to accurately determine the number and specific location of people in the classroom, especially in scenarios where people are stationary, seating is unevenly distributed, or there are few or no people in certain areas.

[0003] In actual teaching scenarios, classrooms often present situations where, for example, lights remain on even after people have left; a few people are using the lighting equipment, yet the entire area is illuminated; and the distribution of students between rows is uneven, but the lights are all on simultaneously. Traditional infrared or microwave detection methods have limitations. For instance, classroom lighting systems have relatively coarse detection granularity, making it difficult to distinguish the number of people, pinpoint seating areas, and are easily affected by stationary individuals. Therefore, it is difficult to achieve the goal of zoned lighting based on the location of people and more precise on-demand control of the lighting system.

[0004] Current classroom lighting solutions based on cameras or image recognition mainly focus on total headcount or simple on / off control, lacking a closed-loop decision-making mechanism that integrates headcount detection results with classroom space zoning, headcount levels, time windows, minimum operating intervals, and forced light-off strategies. Relying solely on whether someone is controlling all the lights leads to wasted localized lighting. Conversely, controlling the lights based on a fixed schedule is ill-suited to various usage patterns such as self-study, exams, projection, and evening study sessions.

[0005] Therefore, it is necessary to propose an intelligent adjustment scheme for the refined management of classroom lighting. By using edge image detection to obtain the location and number of people, the classroom is divided into multiple lighting zones. A mapping relationship between the location of people, the number of people in each zone, and the brightness of the lamps is established. Combined with time constraints and safety control strategies, the classroom lighting control effect of "lights on when people are present, lights off when people leave, dimming on demand, and zoned adjustment" can be achieved. Summary of the Invention

[0006] To address the problems of existing classroom lighting control systems, which struggle to simultaneously sense personnel location and the number of people in a given area, and cannot adjust brightness levels according to changes in the number of people in different areas, leading to issues such as all lights being on or off simultaneously, wasted local lighting, frequent switching, lights not being turned off in time at night, and untraceable control processes, this invention provides a classroom lighting zoned intelligent adjustment system and method based on personnel location and number of people statistics.

[0007] The intelligent classroom lighting zone adjustment system based on personnel location and headcount includes: an image acquisition layer, an edge recognition layer, a region mapping and statistics layer, a lighting rule decision layer, a lighting control execution layer, and a data management and interaction layer. The image acquisition layer uses cameras to acquire real-time images or video frames within the classroom. The edge recognition layer uses a lightweight target detection model on an edge computing device to specifically identify students' faces or heads in the image. After recognition, it provides the bounding box coordinates, confidence score, and total number of detected targets for each target.

[0008] The area mapping statistics layer maps the center point or bottom center point of the detected target to the corresponding lighting area according to the pre-defined classroom partition boundaries, thereby calculating the real-time number of people in each area. The classroom can be divided into a 3×3 grid area, or it can be divided into any number of polygonal areas according to the actual classroom layout. Each area corresponds to one or more sets of independently controllable lighting fixtures.

[0009] The lighting rule decision layer generates control decisions based on the number of people in each area, the regional lighting rules, the current time period, the mandatory light-off time, the allowed operation time window, and the minimum operation interval. The lighting rules here include a mapping relationship between the number of people and brightness, such as 0 people corresponding to 0% brightness, 1 to 3 people corresponding to 50% brightness, 4 to 7 people corresponding to 80% brightness, and 8 or more people corresponding to 100% brightness.

[0010] The implemented solution establishes a five-stage decision-making process based on lighting rules. Throughout the implementation, conditions are checked sequentially. First, it checks if the mandatory light-off time has arrived; next, it checks if the system is in a sleep period; then, it checks if the system is within the permitted operation window; next, it checks if the minimum operation interval is met; and finally, it determines the brightness level based on the number of people. In this process, if any stage's conditions are not met, the system records the reason for skipping the operation, and the lights remain in their current state. Only when all stage conditions are successfully met will the target brightness be calculated based on the actual number of people in the corresponding area.

[0011] The role of the lighting control execution layer is to translate the target brightness into lighting control commands. Using GPIO, PWM, relays, smart lighting buses, or network control interfaces, it drives the lights in the corresponding areas to perform operations such as turning on, turning off, maintaining brightness, or adjusting brightness. The data management and interaction layer is responsible for storing detection records, detailed information on the number of people in each area, current lighting status, rule configurations, lighting operation logs, and system audit logs. It also provides configuration, monitoring, query, and export functions through local clients or a web interface.

[0012] This invention provides a method for intelligent zoning adjustment of classroom lighting based on personnel location and headcount, comprising the following steps: S1. Capture classroom images or video frames; S2. Perform personnel target detection on the classroom image or video frame at the edge to obtain the location coordinates and number of the detected targets; S3. Map the detected target to the corresponding lighting area according to the preset area boundary, and count the number of people in each lighting area; determine the area belonging point by the center point or bottom center point of the detected target bounding box, and determine whether the area belonging point is located within the preset area polygon by the ray method; S4. Determine the candidate target brightness based on the number of people and the brightness mapping rules for each lighting area; S5. Sequentially execute the forced light-off judgment, sleep time judgment, allowed operation window judgment, minimum operation interval judgment, and number of people rule judgment to generate the final target brightness or skip result; if the current time reaches the forced light-off time, directly set the target brightness of the corresponding area to 0; if the current time is in the sleep time or not in the allowed operation window, maintain the current light state and record the skip reason; S6. Control the luminaires in the corresponding lighting area according to the final target brightness; S7. Record detection results, number of people in the area, control actions, target brightness, and decision log.

[0013] Furthermore, It also includes an electronic device, which includes a processor, a memory, an image acquisition interface, and a lighting control interface. The memory stores program instructions, which, when executed by the processor, implement the above-described method.

[0014] It also includes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0015] The beneficial effects of this invention are: 1. Improve the accuracy of lighting control: By using personnel target detection and area mapping statistics, it is possible not only to determine whether there are people in the classroom, but also to determine which zone the people are in, thereby achieving the goal of independent dimming according to the zone.

[0016] 2. By using the “number of people-brightness” mapping rule, energy waste can be reduced, and all lights can be turned on at full power when only a few people are using the classroom, thus reducing ineffective lighting and long-term vacant lighting.

[0017] 3. Reducing frequent switching can avoid frequent switching of lights due to short-term fluctuations in the number of people, thereby improving the lifespan of the lights and the stability of control. This is achieved by using minimum operating intervals and time window constraints.

[0018] 4. Improve nighttime safety management: By using methods such as forced light shutdown and sleep time judgment, lights can be automatically turned off or kept off during nighttime or non-open periods, thereby reducing the cost of manual inspections.

[0019] 5. To enhance traceability and maintainability, detection records, number of people in the area, target brightness, decision results, reasons for skipping, and operation logs are stored together. This facilitates subsequent energy consumption analysis, troubleshooting, and rule optimization. Attached Figure Description

[0020] Figure 1 This is a diagram of the overall system architecture of the present invention; Figure 2 This is a flowchart of the detection and zone dimming process for the present invention. Figure 3 This is a schematic diagram of the partitioned number of people brightness mapping and the five-stage rule engine of the present invention; Figure 4 This is a schematic diagram of the database and log traceability structure of the present invention. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. The following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0022] from Figure 1 As can be seen, this system adopts a layered architecture, consisting of, from top to bottom, an image acquisition layer, an edge recognition layer, a region mapping and statistics layer, a lighting rule decision layer, a lighting control execution layer, and a data management and interaction layer. The hardware for the image acquisition layer can utilize various image acquisition devices such as USB cameras, CSI interface cameras, and network cameras, which can be deployed in suitable locations at the front or back of the classroom to continuously acquire panoramic video frames within the classroom.

[0023] The edge recognition layer can be implemented using Raspberry Pi, embedded industrial PCs, edge AI boxes, or other edge devices with local inference capabilities. In this embodiment, the edge device is a Raspberry Pi 5, and the detection model can be a YOLO series (such as YOLOv8n) or other lightweight target detection models exported in ONNX format for deployment. The aforementioned device models and model types are merely one implementation method. This invention focuses on protecting the mechanisms for personnel location detection, zoning personnel statistics, and closed-loop control of lighting rules, and is not limited to specific hardware or specific models.

[0024] The edge detection layer first scales, normalizes, converts the format of, and performs necessary lighting enhancement on the acquired video frames before inputting them into the object detection model. The object detection model outputs the bounding box coordinates and confidence score of each student's face or head target. The system filters out low-confidence targets based on a pre-set confidence threshold, thus generating detection results that include the target number, bounding box coordinates, confidence score, and acquisition time.

[0025] To adapt to classroom lighting control scenarios, the system uses facial targets, head targets, or upper body targets as detection objects. If students are looking down, have their faces turned to the side, or are slightly obscured, head detection or multi-class detection can be used to improve detection stability. The model inference process is completed locally at the edge, without uploading the original images to the cloud. This reduces reliance on the network and, to some extent, mitigates privacy risks.

[0026] The area mapping statistics layer determines target affiliation based on classroom area boundaries pre-defined by the administrator. Area boundaries can be rectangles, quadrilaterals, or arbitrary polygons. For each detected target, the system can take the geometric center point or bottom center point of the bounding box as the area affiliation point, and use ray casting, a point-within-polygon determination algorithm, or a coordinate range determination algorithm to determine the area to which the target belongs.

[0027] In one implementation, the classroom is divided into nine 3x3 grid areas. Area 0 corresponds to the podium area, while areas 1 through 8 correspond to different seating areas. The system iterates through all detected targets, counts the number of people in each area and the total number of people in the classroom, and then writes the results into a detection record table and an area population details table.

[0028] from Figure 3It can be seen that the lighting rule decision layer configures a "number of people - brightness" mapping rule for each area. Each such rule includes multiple brightness levels, and each brightness level includes information such as minimum number of people, maximum number of people, and target brightness percentage. The system can set different rules for the podium area, front row area, middle row area, and back row area. Of course, it can also combine multiple areas into a rule set and then apply it uniformly to a single camera position or a classroom.

[0029] This rule engine proceeds through five stages in sequence. First, it checks if the current time has reached the mandatory light-off period; if so, it directly outputs the light-off command. Second, it checks if the area is currently in a sleep period; if so, no further control is executed. Third, it determines if the current time falls within the permitted operation window; if not, it records a "non-permitted operation time" and skips all subsequent operations. Fourth, it checks if the minimum operation interval has been met since the last light operation in this area; if the interval is insufficient, it terminates immediately. Fifth, based on the current number of people in the area, it matches the corresponding brightness level and outputs the target brightness.

[0030] The lighting control execution layer receives the control results output by the lighting rule decision layer, and then converts the target brightness percentage into a PWM duty cycle, relay switch signal, or smart lighting protocol instruction. If the LED light fixture supports dimming, the system will output the corresponding PWM signal according to the target brightness. If the light fixture only supports on / off control, the system will treat a target brightness greater than 0 as the light being on, and a target brightness equal to 0 as the light being off.

[0031] After the task is completed, the system updates the current lighting status table, recording the area number, target brightness, current brightness, executed action, execution time, and device feedback results. If the device is offline, GPIO initialization fails, or the lighting operation fails, the system generates an exception event and writes it to the system log.

[0032] In terms of data management and interaction, this function is responsible for saving detection records, detailed information on the number of people in each area, lighting rules, rule set information, current lighting status details, lighting operation logs, manual control logs, and system audit logs. Administrators can use local clients or web interfaces to view real-time population data, area heat dissipation, lighting status details, historical detection records, energy consumption statistics, and information related to abnormal events.

[0033] In automatic control mode, lighting control is triggered by the rules engine; once switched to manual mode, administrators can directly turn lights on or off or adjust brightness in a specific area. For sensitive operations, the system will first verify identity and record the operator, operation time, and the state before and after the operation for easy traceability later.

[0034] The system can set corresponding lighting modes for various scenarios, including classroom learning, self-study periods, projection displays, exams, and all-off modes. Different modes can enable different area brightness rules, forced light-off measures, and time windows. For example, projection mode can reduce the brightness of a portion of the area in front of the podium, exam mode can improve overall brightness uniformity, and all-off mode can directly set the target brightness of all areas to 0.

[0035] The system can further optimize the rules based on timetables, holidays, ambient light sensor data, and classroom reservation information. The aforementioned extended information, as preferred input, will not affect the core technical solution of this invention: zoned lighting adjustment based on personnel location and headcount.

[0036] In one implementation, the system can set the following lighting operation status field to record the decision results and execution status of each area after each detection.

[0037]

Claims

1. A classroom lighting zoning intelligent adjustment system based on personnel position and number statistics, characterized in that, It includes an image acquisition layer, an edge recognition layer, a region mapping and statistics layer, a lighting rule decision layer, a lighting control execution layer, and a data management and interaction layer, among which: The image acquisition layer is responsible for acquiring classroom images or video frames; The edge recognition layer detects people in classroom images or video frames and outputs the location coordinates, confidence level, and number of detected targets. The people detection includes at least one of face detection, head detection, or upper body target detection. The edge recognition layer performs model inference locally on the edge computing device. The region mapping statistics layer maps the detected targets to the corresponding lighting areas based on the preset region boundaries, and counts the number of people in each lighting area; The lighting rule decision layer generates the target brightness according to the number of people in each lighting area, the brightness mapping rule for people, and time constraints. The lighting control execution layer controls the lights in the corresponding lighting area based on the target brightness. The data management and interaction layer is used to save detection records, number of people in the area, lighting status, rule configuration, and lighting operation logs.

2. The system of claim 1, wherein, The region mapping statistics layer divides the classroom into a 3×3 grid area, or into multiple rectangular, quadrilateral, or polygonal areas according to the classroom layout. The region mapping statistics layer determines the region affiliation point based on the geometric center point or bottom center point of the detected target bounding box, and determines the lighting area to which the region affiliation point belongs through a point-within-polygon determination algorithm or ray method.

3. The system according to claim 1, characterized in that, The number of people and brightness mapping rule includes at least one number of people level and a corresponding target brightness percentage. Each number of people level includes a minimum number of people, a maximum number of people, and a target brightness.

4. The system according to claim 1, characterized in that, The lighting rule decision layer includes a five-stage rule engine, which sequentially performs judgments on forced light-off, sleep time, allowed operation window, minimum operation interval, and number of people rule. When any stage of the five-stage rule engine fails to meet the execution conditions, the system maintains the current lighting state and records the reason for skipping. When the execution conditions are met, the system matches the target brightness according to the number of people in the corresponding area and generates control instructions.

5. The system according to claim 1, characterized in that, The lighting control execution layer controls the lighting fixtures in the corresponding area through at least one of the following methods: GPIO, PWM, relay, smart lighting bus, or network control interface.

6. The system according to claim 1, characterized in that, The data management and interaction layer stores the region number, number of people in the region, target brightness, current brightness, action to be performed, reason for skipping, rule identifier, and execution time for each lighting decision.

7. The system according to claim 1, characterized in that, It also has a lighting mode management module, which covers at least one of the following lighting modes: class mode, self-study mode, projection mode, exam mode, and all off mode. Different lighting modes will be bound to different area brightness rules or time constraint parameters.

8. A method for intelligent zoning adjustment of classroom lighting based on personnel location and headcount, characterized in that: Includes the following steps: S1. Capture classroom images or video frames; S2. Perform personnel target detection on the classroom image or video frame at the edge to obtain the location coordinates and number of detected targets; S3. Map the detected target to the corresponding lighting area according to the preset area boundary, and count the number of people in each lighting area; determine the area belonging point by the center point or bottom center point of the detected target bounding box, and determine whether the area belonging point is located within the preset area polygon by the ray method; S4. Determine the candidate target brightness based on the number of people and the brightness mapping rules for each lighting area; S5. Sequentially execute the forced light-off judgment, sleep time judgment, allowed operation window judgment, minimum operation interval judgment, and number of people rule judgment to generate the final target brightness or skip result; if the current time reaches the forced light-off time, directly set the target brightness of the corresponding area to 0; if the current time is in the sleep time or not in the allowed operation window, maintain the current light state and record the skip reason; S6. Control the luminaires in the corresponding lighting area according to the final target brightness; S7. Record detection results, number of people in the area, control actions, target brightness, and decision log.

9. The method according to claim 8, characterized in that: It also includes an electronic device, which includes a processor, a memory, an image acquisition interface, and a lighting control interface. The memory stores program instructions, which, when executed by the processor, implement the method of claim 8.

10. The method according to claim 8, characterized in that: It also includes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method of claim 8.

Images