LED lamp adaptive dimming method and system based on deep learning

Abstract

The present application relates to the technical fields of intelligent lighting control, in particular to a LED lamp adaptive dimming method and system based on deep learning, comprising the following steps: obtaining sampling point illuminance and coordinates to generate a distribution matrix, matching illuminance differences and coverable light sources to form an adjustment list, extracting duty cycle data to compare power and current limits to generate adjustment instructions, distributing phase delay and combining duty cycle information, and combining power and direction data to generate self-matching dimming instructions.In the present application, the spatial mapping is formed by obtaining the relationship between multi-point illuminance and position, the output is adjusted by matching the illuminance difference area and the coverable light source, the brightness is kept stable by comparing the duty cycle, power and current limits, the coordination rhythm is formed by distributing the phase delay and combining the duty cycle, the light source is coordinated in time and intensity, the real-time power and illumination direction are corrected to reduce shadows and glare, thereby improving the lighting balance, visual comfort and energy utilization.

Description

Technical Field

[0001] This invention relates to the field of intelligent lighting control technology, and in particular to an adaptive dimming method and system for LED lights based on deep learning. Background Technology

[0002] The field of intelligent lighting control technology centers on the automated control and management of lighting systems. It systematically integrates sensors, a central controller, communication networks, and lighting actuators such as LEDs. By collecting real-time information such as ambient light intensity, human activity status, or preset schedules, it precisely adjusts the brightness and color temperature of one or more lighting units according to specific control strategies. Traditional LED adaptive dimming, for example, addresses the lighting needs of specific areas by using photosensitive sensors to monitor ambient light intensity in real time or infrared sensors to monitor human activity. This sensor data is input to a microcontroller, which calculates based on preset threshold logic or a proportional-integral-derivative (PID) control algorithm, ultimately outputting a pulse-width modulation (PWM) signal to directly control the LED driver circuit, thereby adjusting the light output intensity.

[0003] Existing technologies in lighting control mostly rely on single-point illumination acquisition combined with a single adjustment signal for processing, lacking the ability to perform correlation analysis based on the overall illuminance distribution of the area. When the illumination demand varies greatly in different locations, brightness gaps or overlaps between adjacent areas can easily occur, leading to glare. In scenarios where illumination demand changes rapidly, due to the lack of phase and intensity coordination between adjustment signals, some light sources may experience fluctuating output brightness or response delays, resulting in unstable lighting effects, decreased visual comfort, and energy imbalance. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide an LED lamp adaptive dimming method based on deep learning, comprising the following steps:

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an LED lamp adaptive dimming method based on deep learning, comprising the following steps:

[0006] S1: Obtain the illuminance information and spatial coordinate data of the LED lighting nodes at the sampling points in the task area, the relationship between the corresponding illuminance information and the PWM control drive power output value, and arrange the distribution matrix according to the spatial mapping order to obtain the illuminance distribution information of the task area.

[0007] S2: Based on the illuminance distribution information of the task area, compare the illuminance positions of the sampling points, match the sampling points with inconsistent target illuminance with the corresponding LED lighting nodes that can be covered, arrange them into a dataset according to the coverage ratio, and obtain the LED node adjustment list.

[0008] S3: Based on the LED node adjustment list, extract the LED lighting node drive control duty cycle data, match the node with the original duty cycle value, and compare it with the power limit and current limit to obtain the node duty cycle adjustment command;

[0009] S4: Based on the node duty cycle adjustment command, call the LED lighting node PWM cycle data, allocate phase delay values ​​according to the number of cycles, and combine them with the duty cycle adjustment command according to the node number to obtain node phase and duty cycle control information;

[0010] S5: Based on the node phase and duty cycle control information, call the real-time power limit and illumination direction data of the LED lighting node, and obtain the LED lamp self-matching dimming control command by corresponding control information and power and direction data.

[0011] As a further embodiment of the present invention, the task area illuminance distribution information includes sampling point illuminance values, node spatial position relationships, and illuminance distribution matrix; the LED node adjustment list includes node identification information, coverage area data, and matching coverage ratio; the node duty cycle adjustment instruction includes node identification information, duty cycle change data, and power and current limiting parameters; the node phase and duty cycle control information includes node number information, phase delay data, and duty cycle control data; and the LED lamp self-matching dimming control instruction includes full node control information, power allocation data, and illumination direction parameters.

[0012] As a further aspect of the present invention, the sampling point illuminance location refers to extracting the spatial coordinates and corresponding illuminance values ​​of the measurement points during the illuminance measurement process in the task area, and inputting them into the matrix structure according to the spatial mapping order to analyze the light distribution and difference range of the differentiated locations.

[0013] The term "coverable LED lighting node" refers to the process of calling the illumination coverage data of LED light nodes during the illuminance difference area processing, spatially comparing the coordinates of the difference area with the illumination range of the nodes, calculating the coverage ratio, and selecting nodes that can provide illumination.

[0014] As a further aspect of the present invention, the duty cycle of the LED lighting node drive control refers to the proportion of the high-level time of the PWM drive signal of each LED node to the entire cycle during the brightness adjustment process. The proportion is compared with the upper limit of the node's power and the current limit to analyze the brightness adjustment space.

[0015] The duty cycle adjustment command refers to correcting the original duty cycle value according to a preset adjustment rule after comparing the duty cycle with the power and current, and then matching the corrected value with the node identifier and control parameters.

[0016] As a further aspect of the present invention, the specific steps of S1 are as follows:

[0017] S101: Obtain the illuminance information and node spatial coordinate data of the LED lighting node at the sampling point in the task area, and input the sampling point illuminance information and PWM control drive power output value into the data table according to the sampling point correspondence to obtain the illuminance and power corresponding data.

[0018] S102: Based on the illuminance and power correspondence data, call the node spatial coordinate data, spatially correspond the node position with the sampling point position, and associate each sampling point with illuminance information and power information to obtain spatially associated illuminance information;

[0019] S103: Based on the spatial correlation illuminance information, the illuminance information under the mapping order is input into the matrix structure according to the row and column positions according to the spatial mapping order of nodes and sampling points to obtain the illuminance distribution information of the task area.

[0020] As a further aspect of the present invention, the specific steps of S2 are as follows:

[0021] S201: Based on the illuminance distribution information of the task area, call the task area coordinates and the target illuminance requirement value, compare the illuminance of the sampling points with the corresponding positions of the target illuminance, divide the inconsistent sampling point position ranges, and obtain the illuminance difference positions.

[0022] S202: Based on the location of the illuminance difference, call the coverage information of the LED lighting nodes, match the location of the illuminance difference with the nodes that can be covered, calculate the coverage ratio of the nodes within the range, filter the set of nodes that can be covered, and obtain the set of covered nodes.

[0023] S203: Based on the set of coverage nodes, call the coverage ratio parameter between nodes and sampling points, allocate the ratio parameter according to the correspondence between nodes and sampling points, and obtain the LED node adjustment list.

[0024] As a further aspect of the present invention, the specific steps of S3 are as follows:

[0025] S301: Based on the LED node adjustment list, call the LED lighting node drive control duty cycle data, match the nodes with the original duty cycle values ​​according to the node identifier, associate the drive channel number with the node identifier, map the relationship, set the index relationship according to the node sequence, and obtain the duty cycle matching result;

[0026] S302: Based on the duty cycle matching result, compare the duty cycle matching result with the power limit and current limit, calculate the current utilization rate of the node, and match the result with the node identifier to obtain the current utilization result;

[0027] S303: Based on the current utilization results, compare the current utilization rate with the duty cycle matching data, replace the corresponding values, and convert them into a control command format that can be directly issued to obtain the node duty cycle adjustment command.

[0028] As a further aspect of the present invention, the specific steps of S4 are as follows:

[0029] S401: Based on the node duty cycle adjustment command, extract the PWM period data of the LED lighting node, divide the node period length into segments according to the number, and correspond them with the node number to obtain the period segment information;

[0030] S402: Based on the periodic segment information, phase delay values ​​are sequentially assigned to the segments, and each phase delay is associated with the corresponding node number to obtain the phase delay assignment result;

[0031] S403: Based on the phase delay allocation result and the node duty cycle adjustment instruction, the phase delay and duty cycle data are sequentially associated according to the node number, and an executable instruction structure is output to obtain the node phase and duty cycle control information.

[0032] As a further aspect of the present invention, the specific steps of S5 are as follows:

[0033] S501: Based on the node phase and duty cycle control information, call the real-time power upper limit data of the LED lighting node, associate the power upper limit with the corresponding node number, and match it with the control information to obtain the power matching result;

[0034] S502: Based on the power matching result, the direction data is sequentially matched with the node number in the power matching result according to the LED lighting node illumination direction data to obtain the direction power matching result;

[0035] S503: Based on the direction power correspondence result, the node phase and duty cycle control information is called, and the node phase, duty cycle, power and direction data are sequentially matched and converted into executable full node control content to obtain the LED lamp self-matching dimming control command.

[0036] A deep learning-based adaptive dimming system for LED lights includes:

[0037] The illuminance data acquisition module acquires the illuminance information and node spatial coordinate data of the LED lighting nodes at the sampling points in the task area. It processes the illuminance information and the PWM control drive power output value according to the correspondence of the sampling points. Through the spatial mapping order of the nodes and sampling points, a distribution matrix is ​​constructed to obtain the illuminance distribution information of the task area.

[0038] Based on the illuminance distribution information of the task area, the task area matching module calls the task area coordinates and the target illuminance requirement value, compares the position of the illuminance of the sampling points, matches the sampling points with insufficient illuminance with the corresponding LED lighting nodes that can be covered, and combines them according to the coverage ratio of the nodes and sampling points to obtain the LED node adjustment list.

[0039] The duty cycle adjustment module, based on the LED node adjustment list, calls the LED lighting node drive control duty cycle data, matches the nodes with the original duty cycle values ​​according to the node identifier, compares the matching results with the node power limit and current limit, and obtains the node duty cycle adjustment command.

[0040] Based on the node duty cycle adjustment command, the phase control module calls the PWM cycle data of the LED lighting node, divides the node cycle length by quantity and allocates the phase delay value, and combines the phase delay value with the corresponding duty cycle adjustment command in the order of node number to form control information, thereby obtaining the node phase and duty cycle control information.

[0041] The dimming command generation module, based on the node phase and duty cycle control information, calls the real-time power limit and illumination direction data of the LED lighting node, corresponds the control information with the power and direction data, and converts it into full-node control content that can be issued and executed, thus obtaining the LED lamp self-matching dimming control command.

[0042] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0043] In this invention, a continuous distribution state is formed by spatial mapping of the illuminance and positional relationship of multiple points. The output is adjusted centrally by matching the illuminance difference area with the coverable light source. The brightness is kept stable by comparing the duty cycle with the power and current limits. The phase delay is allocated and combined with the duty cycle to form a coordinated rhythm, so that the light source is coordinated in time and intensity. The shadows and glare are reduced by combining real-time power and illumination direction correction, thereby improving the lighting uniformity, visual comfort and energy consumption efficiency. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of the steps of the present invention;

[0046] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0047] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0048] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0049] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0050] Figure 6 This is a detailed schematic diagram of S5 of the present invention;

[0051] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0052] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0053] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0054] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0055] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0056] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0057] Please see Figure 1 This invention provides a deep learning-based adaptive dimming method for LED lights, comprising the following steps:

[0058] S1: Obtain the illuminance information and spatial coordinate data of the LED lighting nodes at the sampling points within the task area. Process the illuminance information and the power output value driven by PWM control according to the correspondence of the sampling points. Construct a distribution matrix through the spatial mapping order of the nodes and sampling points to obtain the illuminance distribution information of the task area.

[0059] S2: Based on the illuminance distribution information of the task area, call the task area coordinates and the target illuminance requirement value, compare the position of the illuminance of the sampling points, match the sampling points with insufficient illuminance with the LED lighting nodes that can be covered, and combine them according to the coverage ratio of the nodes and sampling points to obtain the LED node adjustment list.

[0060] S3: Based on the LED node adjustment list, call the drive control duty cycle data of the LED lighting nodes, match the nodes with the original duty cycle values ​​according to the node identifier, compare the matching results with the node power limit and current limit, and obtain the node duty cycle adjustment command.

[0061] S4: Based on the node duty cycle adjustment command, call the PWM cycle data of the LED lighting node, divide the node cycle length according to the number and then allocate the phase delay value in sequence, and combine the phase delay value with the corresponding duty cycle adjustment command in the order of node number to form control information, so as to obtain the node phase and duty cycle control information.

[0062] S5: Based on the node phase and duty cycle control information, call the real-time power limit and illumination direction data of the LED lighting node, correspond the control information with the power and direction data, convert it into full node control content that can be issued and executed, and obtain the LED lamp self-matching dimming control command.

[0063] The task area illuminance distribution information includes illuminance values ​​at sampling points, spatial relationship of nodes, and illuminance distribution matrix. The LED node adjustment list includes node identification information, coverage area data, and matching coverage ratio. The node duty cycle adjustment instruction includes node identification information, duty cycle change data, and power and current limit parameters. The node phase and duty cycle control information includes node number information, phase delay data, and duty cycle control data. The LED lamp self-matching dimming control instruction includes full node control information, power distribution data, and illumination direction parameters.

[0064] Please see Figure 2 The specific steps of S1 are as follows:

[0065] S101: Obtain the illuminance information and node spatial coordinate data of the LED lighting node at the sampling point in the task area, and input the sampling point illuminance information and PWM control drive power output value into the data table according to the sampling point correspondence to obtain the illuminance and power corresponding data.

[0066] First, the sampling order is determined based on the fixed spatial location of each sampling point within the task area. Each sampling point is then placed sequentially within the sensing range of its corresponding light sensor to collect illuminance readings. For example, 25 sampling points are evenly distributed within a 5m x 5m area, with a 1m interval between each point. The illuminance value corresponding to each location is recorded sequentially. Subsequently, when collecting the spatial coordinate data of the nodes, the precise coordinates of each LED lighting node within the area must be determined using measuring instruments. For example, node A is located at (2.0, 3.0), and node B is located at (4.0, 1.0). These coordinate values ​​are used for subsequent positioning and correlation of illuminance information. When correlating the illuminance information of the sampling points with the PWM control drive power output value, the illuminance data of each sampling point needs to be paired with the PWM drive output power of the nearest LED node. For example, the illuminance of sampling point P1 is 32 lux. 0. The PWM output of the corresponding node A is 75%, while the illuminance of sampling point P2 is 280, and the PWM output of the corresponding node B is 60%. This pairing needs to be completed point by point to ensure no omissions or misalignments. Then, these one-to-one illuminance and power pairing data are entered into a data table. Each row in the data table corresponds to a sampling point, including the sampling point number, illuminance value, corresponding node number and its PWM output power. When entering the data, attention should be paid to the uniformity of numerical units, such as illuminance in lx and power output in percentage, and the data should be filled in according to the sampling point number order. In a real scenario, for example, if there are 10 LED nodes and 20 sampling points in an indoor area, 20 records need to be entered into the data table. Each record fully reflects the correspondence between the illuminance of the sampling point and the power output of its corresponding node, thus obtaining the illuminance and power corresponding data.

[0067] S102: Based on the illuminance and power correspondence data, call the node spatial coordinate data, spatially correspond the node position with the sampling point position, and associate each sampling point with illuminance information and power information to obtain spatially associated illuminance information;

[0068] First, extract the illuminance value and power output percentage corresponding to each sampling point, ensuring the uniqueness of each sampling point number. Then, retrieve the node spatial coordinate data to extract the planar position coordinates of each LED lighting node. For example, the coordinates of node N1 are (1.5, 2.0), and the coordinates of node N2 are (3.0, 4.0). Next, spatially map the node positions to the sampling point positions. This can be done by comparing the coordinate differences one by one, associating the node position with the sampling point position with the smallest distance. For example, the coordinates of sampling point P5 are (1.6, 2.1), which corresponds to node N1. After completing the spatial mapping, compare the illuminance value of 320 lx obtained from the sampling point with the coordinates of node N1. The power output value of 78% is simultaneously associated with the spatial location information of the sampling point. All sampling points are processed sequentially to ensure that each sampling point is associated with a nearest neighbor node and its power information. During this process, if two nodes are equidistant from the same sampling point, the node with the smaller number is selected for association according to the order of node numbers to avoid duplicate allocation of sampling points. For example, if sampling point P8 is 0.5 meters away from both nodes N3 and N4, then node N3 with the smaller number is selected as the corresponding node. After the above point-by-point matching and information association operations, the illuminance information and power information of all sampling points are mapped to their associated nodes according to their spatial location to obtain spatially associated illuminance information.

[0069] S103: Based on spatial correlation illuminance information, the illuminance information under the mapping order is input into the matrix structure according to the row and column positions according to the spatial mapping order of nodes and sampling points to obtain the illuminance distribution information of the task area.

[0070] First, extract the number and corresponding illuminance value of each sampling point in the spatial mapping order. For example, sampling point P1 corresponds to an illuminance of 285 lx, and P2 corresponds to an illuminance of 310 lx. Preserve their spatial association with the nodes. Then, according to the spatial mapping order between nodes and sampling points, arrange the sequential positions of the sampling points in row and column order. At this point, the matrix row and column range of the task area needs to be set. For example, the area width is divided into 4 columns, and the height into 3 rows. During allocation, fill the sampling point with the smallest number sequentially into the upper left starting position of the matrix, and fill in the numbers sequentially in ascending order of rows. For example, the first row and first column is P1, the first row and second column is P2, and so on. After one row is filled, move on to the next row. When inputting illuminance information, the illuminance value of each sampling point needs to be directly entered. Fill in the corresponding positions in the matrix. For example, fill in 285 in the first row and first column, and 310 in the second row and second column. If a sampling point does not have an illuminance value, it is necessary to call the illuminance of its nearest known sampling point to fill it in. The point with the smallest distance in both the horizontal and vertical directions is selected as the reference. For example, if the fourth column of the first row is missing a value, the illuminance value of the sampling point in the third column of the first row or the fourth column of the second row is used to fill it in. After all rows and columns are filled, the matrix needs to be checked as a whole. Check whether the illuminance information of the sampling points in each row and column is consistent with the original spatial mapping order. If the order is found to be incorrect, adjust the illuminance value and number relationship of the corresponding position until all matrix positions are completely consistent with the spatial mapping order to obtain the illuminance distribution information of the task area.

[0071] Please see Figure 3 The specific steps of S2 are as follows:

[0072] S201: Based on the illuminance distribution information of the task area, call the coordinates of the task area and the target illuminance requirement value, compare the illuminance of the sampling points with the corresponding positions of the target illuminance, divide the range of inconsistent sampling point positions, and obtain the illuminance difference positions.

[0073] First, extract the spatial coordinates and corresponding illuminance values ​​of each sampling point in the area. For example, sampling point P1 is located at area coordinates (2.5, 4.0) with an illuminance of 285 lx, and sampling point P2 is located at (3.0, 4.0) with an illuminance of 310 lx. Then, call the target illuminance requirement value for the task area. These requirement values ​​are usually pre-planned according to the purpose of the area. For example, the target illuminance for each coordinate point in the same work area is 300 lx. Based on this, compare the illuminance of each sampling point with the target illuminance corresponding to its coordinates one-to-one to determine whether the illuminance value is equal to the target illuminance. If the illuminance value is greater than the target value, it is recorded as a point that is too high. If the illuminance value is less than the target value, it is recorded as a low point; if the illuminance is equal, it is marked as a matching point. During the comparison process, all sampling points should be compared one by one to ensure full coverage. For example, if the illuminance of P1 is 285 lx, which is lower than the target value of 300 lx, it is judged as a low point; if the illuminance of P2 is 310 lx, which is higher than the target value of 300 lx, it is judged as a high point. After comparing all sampling points, the locations of sampling points with inconsistent illuminance need to be spatially aggregated according to coordinates to find the continuous or adjacent distribution range of these points in the area. For example, when the distance between adjacent coordinate points in the horizontal or vertical direction is less than 0.5 meters, they can be classified as the same difference location range to obtain the illuminance difference location.

[0074] S202: Based on the location of illuminance difference, call the coverage information of LED lighting nodes, match the location of illuminance difference with the nodes that can be covered, calculate the coverage ratio of the nodes within the range, filter the set of nodes that can be covered, and obtain the set of covered nodes.

[0075] The specific formula for calculating the coverage ratio of a node within its range is as follows:

[0076]

[0077] Among them, κ ij S represents the coverage ratio within the range of the illuminance difference at location j of LED lighting node i. ij γ represents the actual coverage area at position j representing the illuminance difference of LED lighting node i. i S represents the dimensionless correction coefficient for the lighting efficiency of LED lighting node i. total δ represents the total target coverage area at location j where illuminance differs. max The standard maximum coverage distance for LED lighting nodes, d ij α represents the straight-line distance between LED lighting node i and the center point of the illuminance difference location j, n represents the total number of neighboring nodes that have a coverage intersection with the illuminance difference location j, k represents the loop index variable of neighboring nodes in the summation operation, and α represents the linear distance between LED lighting node i and the center point of the illuminance difference location j. k O represents the dimensionless influence factor of the coverage overlap of the k-th neighboring node. ikβ represents the overlapping coverage area between LED lighting node i and its k-th neighboring node. k d represents the dimensionless weighting factor representing the distance to the k-th neighboring node. kj This represents the straight-line distance between the k-th adjacent node and the center point of the illuminance difference location j;

[0078] The formula's calculation logic is as follows: First, the basic ratio of the node coverage area to the total target area is calculated through division. Then, lighting efficiency correction is introduced through multiplication. Next, the distance attenuation effect is calculated through subtraction and division, and the distance weighting effect is enhanced through exponentiation. Then, the overlap effect of all adjacent nodes is accumulated through summation. Inside the summation, subtraction is used to compare the difference between the area overlap ratio and the distance weight ratio. The absolute value is positive, and the coverage ratio is obtained through square root operation. The calculation method can take into account multiple influencing factors such as coverage area, distance attenuation, and node overlap.

[0079] Actual measurements showed that:

[0080] Node 1 has a coverage area of ​​S 11 = 285.6 square meters;

[0081] Node 2 covers an area of ​​S 21 = 312.4 square meters;

[0082] Node 3 covers an area of ​​S 31 = 278.9 square meters;

[0083] The measured luminous flux of node 1 is 18,000 lumens, and the standard luminous flux is 24,000 lumens.

[0084] Calculated

[0085] The measured luminous flux at node 2 is 19680 lumens, and the calculated luminous flux is...

[0086] The measured luminous flux at node 3 is 16320 lumens, and the calculated luminous flux is...

[0087] S total The area of ​​the rectangular region at the location of the illuminance difference is 25 meters long and 15 meters wide, measured by a laser rangefinder.

[0088] Calculate S total =25 × 15 = 375 square meters;

[0089] δ max According to the outdoor lighting design specification in the LED lighting standard GB50034-2013, the height is set at 18 meters.

[0090] d 11 = 8.2 meters, d21 = 12.5 meters, d 31 = 15.1 meters;

[0091] The number of adjacent nodes is determined by spatial topology analysis, n=2;

[0092] The average illuminance measured in the overlapping area of ​​node 1 and its adjacent node k1 is 420 lux, and the individual illumination intensity is 300 lux. Calculate... After attenuation correction, α1 = 0.35. The average illuminance measured in the overlapping area of ​​node 1 and its adjacent node k2 is 480 lux, and the individual illumination intensity is 340 lux. Calculate... After correction, α² = 0.42;

[0093] O ik Through geometric calculations of the overlapping lighting areas, O 11 = 45.8 square meters, O 12 = 52.3 square meters;

[0094] β k The distance between adjacent node k1 and the target location is 10.5 meters, and the distance between adjacent node k2 and the target location is 14.8 meters.

[0095] After adjustment, β1 = 0.65;

[0096] After adjustment, β2 = 0.58;

[0097] d 1j = 10.5 meters, d 2j = 14.8 meters;

[0098] Threshold setting process: The lighting efficiency coefficient threshold range [0.2, 1.0] is set according to the LED lighting industry standard JGJ / T119-2008, where 0.2 corresponds to the minimum acceptable lighting efficiency. The calculation process is 20% of the standard luminous flux, i.e. The minimum operating state corresponding to aging or low-power LEDs is 1.0, which corresponds to full-load lighting efficiency, indicating that the LED reaches the design standard luminous flux. The coverage ratio threshold of 0.6 was determined through experimental verification. In lighting tests of 100 groups of differentiated scenarios, the relationship between coverage ratio and lighting effect satisfaction was statistically analyzed. Nodes with a coverage ratio exceeding 0.6 had a 95.8% probability of providing sufficient lighting effects, and the satisfaction rate was calculated. To meet reliability requirements, the selection threshold is set to 0.6;

[0099] Substitute the parameter values ​​into the formula for calculation, taking node 1 as an example:

[0100]

[0101] First term calculation:

[0102] Distance attenuation term:

[0103] Basic term: 0.571 × 0.296 = 0.169;

[0104] The first term of the summation (k=1):

[0105]

[0106] The second term of the summation (k=2):

[0107]

[0108] The total sum of the terms is: 0.777 + 0.970 = 1.747;

[0109] calculate:

[0110] Calculate node 2 using the same method:

[0111]

[0112] =1.260;

[0113] Compute node 3:

[0114]

[0115] When γ i ∈[0.2,1.0] and κ ij When the value is greater than 0.6, the coverage ratio is transmitted to the LED controller module to trigger the node selection mechanism. Because κ... 11 =1.384>0.6, κ 21 =1.260>0.6, κ 31 =1.127>0.6, and γ of all nodes i All three nodes are within a reasonable range and are selected into the set of covered nodes. The set of covered nodes is then filtered to obtain the set of covered nodes.

[0116] Interpretation and numerical significance of results: The results show that the three LED lighting nodes have strong coverage capabilities at locations with illuminance differences, with coverage ratios of 1.384, 1.260, and 1.127, respectively, all exceeding the baseline threshold of 0.6. Comparison with the preset interval [0.6, 2.0] shows that all values ​​are within a reasonable range, indicating that the nodes can provide sufficient lighting coverage for locations with illuminance differences. The coverage ratio values ​​are directly converted into node selection flags. ijWhen the value is greater than 0.6, the selection flag is set to 1; otherwise, it is 0. This completes the conversion from the coverage ratio value to the coverage node set, which includes node 1, node 2, and node 3.

[0117] Explanation of the innovative aspects of the formula:

[0118] The advantage of the formula lies in the introduction of a lighting efficiency correction factor γ. i The power operation of the distance attenuation factor was used to perform a dual optimization evaluation of the node's lighting capability and spatial location, while also considering the overlap influence factor α. k and distance weighting factor β k The absolute difference calculation eliminates the problems of light interference and resource redundancy between adjacent nodes, and avoids the shortcomings of traditional methods that only refer to a single coverage area and ignore the synergistic effect between nodes.

[0119] S203: Based on the set of coverage nodes, call the coverage ratio parameter of the nodes and sampling points, allocate the ratio parameter according to the correspondence between the nodes and sampling points, and obtain the LED node adjustment list;

[0120] First, obtain the unique ID of each node in the set and the list of sampling point IDs it can cover. For example, node N1 can cover sampling points P1, P2, and P4, and node N2 can cover sampling points P3, P5, and P6. Then, call the stored node-to-sampling-point coverage ratio parameters. These ratio parameters can be preset according to the illumination experiment data. For example, the coverage ratio of node N1 to sampling points P1 is 0.65, P2 is 0.50, and P4 is 0.40, and the coverage ratio of node N2 to sampling points P3 is 0.70, P5 is 0.60, and P6 is 0.55. After calling, the ratio parameters are allocated according to the correspondence between nodes and sampling points, that is, each ratio value is accurately assigned to the corresponding node ID and sampling point ID. In the corresponding items, for example, N1-P1 corresponds to a ratio of 0.65, N1-P2 corresponds to a ratio of 0.50, N1-P4 corresponds to a ratio of 0.40, N2-P3 corresponds to a ratio of 0.70, and so on. During the process, it is necessary to check whether each ratio value is between 0 and 1 and meets the accuracy requirements. If the ratio value is out of range, it needs to be removed or corrected. After the allocation is completed, all nodes and their corresponding items of coverage ratio are arranged in a centralized manner so that each node is associated with all its corresponding ratio information. For example, N1 is associated with 0.65, 0.50, and 0.40, and N2 is associated with 0.70, 0.60, and 0.55. Finally, these allocated ratio information are merged into a list in the order of node number to obtain the LED node adjustment list.

[0121] Please see Figure 4 The specific steps of S3 are as follows:

[0122] S301: Based on the LED node adjustment list, call the LED lighting node driver control duty cycle data, match the nodes with the original duty cycle values ​​according to the node identifier, associate the driver channel number with the node identifier, map the relationship, set the index relationship according to the node sequence, and obtain the duty cycle matching result;

[0123] First, extract the unique identifier number for each LED node in the list, such as N01, N02, and N03. Then, retrieve the corresponding drive control duty cycle data for each number. This data can be directly derived from the original settings stored in the control drive file, for example, N01 is 48%, N02 is 52%, and N03 is 60%. Next, match these node numbers with their corresponding duty cycle values ​​one by one to ensure that each number accurately corresponds to its original duty cycle value. During this process, it is also necessary to verify that the duty cycle value range is between 0 and 100 and the decimal precision does not exceed two digits. If the value is found to be out of range or the precision is incorrect, it needs to be re-checked. After reading or correcting and matching, continue to associate the drive channel number with the node identifier. For example, channel CH01 corresponds to N01, CH02 corresponds to N02, and CH03 corresponds to N03. Then, associate and map the relationship between the drive channel number and the node identifier with the previous duty cycle matching results to ensure that each channel can not only find the corresponding node, but also accurately associate its original duty cycle value. Finally, set an index relationship for all nodes according to the node number order. For example, create indexes 1, 2, and 3 in the order of N01, N02, and N03, so that the node number, channel number, and duty cycle can be quickly located according to the index to obtain the duty cycle matching result.

[0124] S302: Based on the duty cycle matching result, compare the duty cycle matching result with the power limit and current limit, calculate the current utilization rate of the node, and match the result with the node identifier to obtain the current utilization result;

[0125] The specific formula for calculating the current utilization rate of a node is as follows:

[0126]

[0127] in, D represents the current utilization rate of node y. y The duty cycle matching result represents node y. The unit representing the upper limit of power for node y is watts (W), U y ζ represents the terminal voltage of node y at the time of calculation. y The duty cycle current efficiency correction coefficient represents node y. δ represents the current limit at node y. y The current measurement deviation correction amount represents node y;

[0128] D yThe average result of multiple measurements was 0.60.

[0129] The rated power on the equipment nameplate is determined by the design limit, and the value is taken as 50,000W;

[0130] U y Low-voltage power distribution range (230V and 400V), take 400V;

[0131] ζ y Based on the fitting of operational data and experimental calibration, the correction factor value is 0.95.

[0132] The value is determined based on the conductor's current-carrying capacity and the protection device's setting, and is taken as 125A.

[0133] δ y : Obtained through calibration data of the measuring equipment, take 2A;

[0134] Operator logic explanation: Numerator of formula First, convert power to current (divide by voltage U). y Then, a correction coefficient ζ is introduced through multiplication. y The equivalent current value is obtained; compared with the current limit. The subtraction is performed, and the absolute value is taken to indicate that regardless of whether the current exceeds or falls below the limit, the deviation is included based on its magnitude. The denominator is then calculated by taking the square root of the sum of squares, making the corrected current limit deviation and the equivalent current itself scalar quantities, thus forming a normalized ratio to obtain the dimensionless utilization rate.

[0135] Calculate the equivalent current value:

[0136]

[0137] Calculate the numerator value (absolute difference):

[0138] |71.25-125|=53.75A;

[0139] Calculate the denominator (sum of squares and square root):

[0140]

[0141] Calculate utilization rate:

[0142]

[0143] Interpretation of results and numerical significance: The results show that the current utilization rate of node y is 0.369, which directly corresponds to the calculation result of step S302. It can be used as a reference indicator of the node's operating status in the system control strategy, and can be used for horizontal comparison with nodes or as input data for scheduling and current limiting judgment.

[0144] Explanation of the innovative aspects of the formula:

[0145] The advantage of the formula lies in the introduction of a correction coefficient ζ. y and measurement deviation correction δ y The calculations of the original power and current limits are revised to make the calculation results closer to the actual operating state of the node, taking into account the actual measurement errors and load characteristics.

[0146] S303: Based on the current utilization results, compare the current utilization rate with the duty cycle matching data, replace the corresponding values, and convert them into a control command format that can be directly issued to obtain the node duty cycle adjustment command.

[0147] First, the current utilization rate of each LED lighting node under its current operating state is extracted. For example, node A is 0.82, node B is 0.76, and node C is 0.91. These values ​​are confirmed to be calculated as the ratio of the node's real-time current to its allowable current limit. Then, duty cycle matching data is retrieved. This data consists of the node number and the original duty cycle value; for example, node A corresponds to 48%, node B to 52%, and node C to 60%. Next, the current utilization rate is compared one by one with the corresponding node's duty cycle value. The comparison method maps the current utilization rate to the duty cycle adjustment rules. For example, when the current utilization rate is below 0.8, the duty cycle is increased by 5. The duty cycle is adjusted by 3 percentage points, remaining unchanged between 0.8 and 0.9, and decreasing by 3 percentage points if it is higher than 0.9. The duty cycle of each node is replaced according to this rule. For example, node A remains at 48%, node B is adjusted to 57%, and node C is adjusted to 57%. During the replacement process, it is also necessary to confirm that the adjusted duty cycle value does not exceed the range of 0 to 100 and retains two decimal places. Then, the replaced duty cycle data is re-associated with the node number and matched with the control field sequence that can be directly recognized by the drive unit. The number, duty cycle value, and current utilization rate of each node are arranged in a standard format to ensure that they can be directly issued and executed to obtain the node duty cycle adjustment instruction.

[0148] Please see Figure 5 The specific steps of S4 are as follows:

[0149] S401: Based on the node duty cycle adjustment command, extract the PWM cycle data of the LED lighting node, divide the node cycle length into segments according to the number, and correspond them with the node number to obtain the cycle segment information;

[0150] First, extract the PWM cycle data corresponding to each LED lighting node. The cycle data reflects the total time length of the node within a PWM control cycle. For example, the cycle length of node 1 is 1200 counting units, node 2 is 1000 counting units, and node 3 is 800 counting units. Then, determine the number of segments based on the number of nodes. For example, three nodes need to correspond to three segments. Divide the counting range of node 1 from 0 to 1199 into one segment, the cycle length of node 2 from 0 to 999 into another segment, and the cycle length of node 3 from 0 to 799 into another segment. During this process, it is necessary to ensure that the counting range of each segment covers the entire cycle without overlap or gaps. Next, match these cycle segments according to the node number order. For example, node 1 corresponds to segment 0-1199, node 2 corresponds to segment 0-999, and node 3 corresponds to segment 0-799. During matching, confirm that the node number is consistent with the number in the adjustment command. Through this one-to-one correspondence, clarify the time slice position and length occupied by each node in PWM control, and obtain the cycle segment information.

[0151] S402: Based on the periodic segment information, phase delay values ​​are sequentially assigned to the segments, and each phase delay is associated with the corresponding node number to obtain the phase delay assignment result;

[0152] First, determine the number of segments and their corresponding node numbers. For example, the periodic segment information contains three segments: Node 1, Node 2, and Node 3, with counting ranges of 0-1199, 0-999, and 0-799, respectively. Then, assign phase delay values ​​to each segment sequentially, following the order of node numbers from smallest to largest. For example, Node 1 is assigned a delay value of 50, Node 2 is assigned a delay value of 100, and Node 3 is assigned a delay value of 150. This delay value represents the number of delays in the PWM signal startup of that node relative to the initial time. During the allocation process, it must be ensured that the delay value is less than the total length of the corresponding periodic segment and does not overlap with the delay range of other nodes. Next, associate each assigned phase delay value with its corresponding node number to form a matching relationship between nodes and phase delays. For example, Node 1 is matched with a delay value of 50, Node 2 with a delay value of 100, and Node 3 with a delay value of 150. During the association process, confirm that the node number is completely consistent with the node number in the periodic segment information, thereby ensuring that the delay value can be accurately applied to the corresponding node and obtaining the phase delay allocation result.

[0153] S403: Based on the phase delay allocation result and the node duty cycle adjustment instruction, the phase delay and duty cycle data are sequentially associated according to the node number, and an executable instruction structure is output to obtain the node phase and duty cycle control information;

[0154] First, read each node number and its corresponding delay value from the phase delay allocation result sequentially. For example, node A corresponds to a delay value of 40, node B to 85, and node C to 120. Then, read each node number and its duty cycle value from the node duty cycle adjustment command. For example, node A is 65%, node B is 70%, and node C is 60%. Next, match the node numbers in order, mapping the delay values ​​to the duty cycles one-to-one. If the node numbers in the phase delay allocation result and the node numbers in the duty cycle adjustment command are completely identical, a direct correspondence is established. If a node number is missing in one direction... If necessary, the corresponding values ​​need to be supplemented during the matching process to ensure data integrity. After matching the delay values ​​and duty cycle data of all nodes, each set of matched data is associated with the node number. For example, node A corresponds to a delay value of 40 and a duty cycle of 65%, node B corresponds to a delay value of 85 and a duty cycle of 70%, and node C corresponds to a delay value of 120 and a duty cycle of 60%. Finally, the associated data is arranged in the order of fields that can be directly recognized by the controller. The fields include three parts: node number, phase delay value, and duty cycle. The uniqueness of each data in the node number is ensured to obtain the node phase and duty cycle control information.

[0155] Please see Figure 6 The specific steps of S5 are as follows:

[0156] S501: Based on the node phase and duty cycle control information, call the real-time power upper limit data of the LED lighting node, associate the power upper limit with the corresponding node number, and match it with the control information to obtain the power matching result;

[0157] First, the node number, phase delay value, and duty cycle value are read sequentially from the control information. For example, node 1 has a phase delay of 40 and a duty cycle of 65%, node 2 has a phase delay of 85 and a duty cycle of 70%, and node 3 has a phase delay of 120 and a duty cycle of 60%. Then, the real-time power limit data table for LED lighting nodes is called, and the power limit value corresponding to each node number is read from this table. For example, the power limit for node 1 is 18W, for node 2 it is 20W, and for node 3 it is 15W. Next, the node numbers are matched, associating the power limit values ​​with nodes with the same numbers in the control information. During the matching process, node numbers are compared one by one. If a number in the power data table does not exist in the control information, that number is skipped. If a node number in the control information is missing from the power data table, the power limit for that node is supplemented by querying historical records or measured values. After matching, the node number, phase delay, duty cycle, and power limit value of each node are associated side by side. For example, the associated result for node 1 is number 1, phase delay 40, duty cycle 65%, and power limit 18W; the associated result for node 2 is number 2, phase delay 85, duty cycle 70%, and power limit 20W; and the associated result for node 3 is number 3, phase delay 120, duty cycle 60%, and power limit 15W. After the association is completed, the power limit of each set of data is checked to see if it is within a reasonable range. For example, 10W to 25W is defined as a reasonable range. If a node exceeds this range, it needs to be marked for subsequent correction. Finally, the power matching result is obtained.

[0158] S502: Based on the power matching results, the direction data is matched sequentially with the node numbers in the power matching results according to the LED lighting node illumination direction data to obtain the direction power matching results;

[0159] First, the node number and its corresponding power limit, phase delay, and duty cycle information are read line by line from the power matching data. For example, node 1 has a power limit of 18W, a phase delay of 40°, and a duty cycle of 65%; node 2 has a power limit of 20W, a phase delay of 85°, and a duty cycle of 70%; and node 3 has a power limit of 15W, a phase delay of 120°, and a duty cycle of 60%. Then, the illumination direction parameters corresponding to each node number are read from the LED lighting node illumination direction data table. These parameters typically consist of two parts: a horizontal angle and a vertical angle, expressed in degrees. For example, node 1's illumination direction is 90° horizontal and 30° vertical; node 2's is 45° horizontal and 20° vertical; and node 3's is 135° horizontal and 25° vertical. After reading the direction parameters, a matching operation is performed on each node number, comparing the node numbers in the power matching result table with those in the illumination direction data table. When the numbers are the same, the corresponding illumination direction and power matching information are merged to form a single record. During the matching process, the completeness of the direction data needs to be checked. If a node number is missing direction parameters, the horizontal and vertical angles of the node are supplemented by on-site measurement, or the average value of adjacent nodes is used as a substitute. After all nodes are matched, the result should include five items: node number, power limit, phase delay, duty cycle, and illumination direction. For example, the complete record of node 1 is number 1, power limit 18W, phase delay 40, duty cycle 65%, illumination direction 90° horizontal and 30° vertical; node 2 is number 2, power limit 20W, phase delay 85, duty cycle 70%, illumination direction 45° horizontal and 20° vertical; and node 3 is number 3, power limit 15W, phase delay 120, duty cycle 60%, illumination direction 135° horizontal and 25° vertical. After sorting, the direction and power correspondence results are formed.

[0160] S503: Based on the direction power correspondence result, call the node phase and duty cycle control information, and sequentially match the node phase, duty cycle, power and direction data to convert them into executable full node control content, thus obtaining the LED lamp self-matching dimming control command;

[0161] First, each record in the direction power correspondence table is read, and the node number, phase delay value, duty cycle value, power limit value, and illumination direction data are extracted sequentially. This information is input as basic parameters into the combined processing flow. Then, the node phase and duty cycle control information table is called, and each node number is compared. When the numbers in the two tables match exactly, the phase and duty cycle values ​​in the control information table are merged with the power limit value and illumination direction in the direction power correspondence table. During the merging process, the range of phase and duty cycle values ​​needs to be verified. For example, the phase value is limited to 0 to 360 degrees, and the duty cycle value is limited to 0% to 100%. If there are values ​​outside the range, adjustments are made according to the set correction rules. For example, when the node number is 1, the phase delay is 40 degrees, the duty cycle is 65%, the power limit is 18W, and the illumination direction is 90° horizontal and 30° vertical. The complete control information for this node should be... These four types of data are directly integrated to form an executable record. After processing all node numbers, the records are arranged in ascending order of node number, and the arrangement results are indexed to ensure that the control commands can be executed in the order of the nodes when they are issued. For example, the record for node 2 is a phase delay of 85 degrees, duty cycle of 70%, power limit of 20W, and illumination direction of 45° horizontal and 20° vertical. The record for node 3 is a phase delay of 120 degrees, duty cycle of 60%, power limit of 15W, and illumination direction of 135° horizontal and 25° vertical. After integration, all records need to be checked for consistency, such as checking whether there is a conflict in the phase value under the same illumination direction. If there is a conflict, the parameters are fine-tuned according to the principle of prioritizing the higher power limit. After completing all matching and verification, the phase, duty cycle, power, and direction data of these nodes are combined in sequence to form the full node control content that can be directly issued, and finally the self-matching dimming control command for LED lights is obtained.

[0162] Please see Figure 7 A deep learning-based adaptive dimming system for LED lights includes:

[0163] The illuminance data acquisition module acquires the illuminance information and node spatial coordinate data of the LED lighting nodes at the sampling points in the task area. It processes the illuminance information and the PWM control drive power output value according to the correspondence of the sampling points. Through the spatial mapping order of the nodes and sampling points, a distribution matrix is ​​constructed to obtain the illuminance distribution information of the task area.

[0164] The task area matching module uses the task area illuminance distribution information to call the task area coordinates and target illuminance requirement value, compares the illuminance of the sampling points, matches the sampling points with insufficient illuminance with the corresponding LED lighting nodes that can be covered, and combines them according to the coverage ratio of nodes and sampling points to obtain an LED node adjustment list.

[0165] The duty cycle adjustment module is based on the LED node adjustment list. It calls the duty cycle data of the LED lighting node driver control, matches the nodes with the original duty cycle values ​​according to the node identifier, and compares the matching results with the node power limit and current limit to obtain the node duty cycle adjustment command.

[0166] The phase control module calls the PWM cycle data of the LED lighting nodes based on the node duty cycle adjustment command, divides the node cycle length according to the number and allocates the phase delay value, and combines the phase delay value with the corresponding duty cycle adjustment command in the order of node number to form control information, thus obtaining the node phase and duty cycle control information.

[0167] The dimming command generation module, based on the node phase and duty cycle control information, calls the real-time power limit and illumination direction data of the LED lighting node, maps the control information to the power and direction data, and converts it into full-node control content that can be issued and executed, thus obtaining the LED lamp self-matching dimming control command.

[0168] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A deep learning-based LED lamp adaptive dimming method, characterized in that, Comprise the following steps: S1: obtaining the LED lighting node in the task area sampling point illuminance information and node space coordinate data, corresponding illuminance information and PWM control driving power output value relationship, according to the space mapping sequence arrangement distribution matrix, the task area illuminance distribution information is obtained; S2: based on the task area illuminance distribution information, the sampling point illuminance position is compared, the sampling point with target illuminance inconsistency is matched with the coverable LED lighting node, and the data set is arranged according to the coverage ratio, and the LED node adjustment list is obtained; S3: based on the LED node adjustment list, the LED lighting node driving control duty cycle data is extracted, the node and the original duty cycle value are matched, and the power upper limit and the current limit value are compared, and the node duty cycle adjustment instruction is obtained; S4: based on the node duty cycle adjustment instruction, the LED lighting node PWM cycle data is called, the cycle is distributed according to the number of phase delay values, and the node number combination is arranged according to the duty cycle adjustment instruction, and the node phase and duty cycle control information is obtained; S5: based on the node phase and duty cycle control information, the LED lighting node real-time power upper limit and illumination direction data are called, the control information is matched with the power and direction data, and the LED lamp self-matching dimming control instruction is obtained; The sampling point illuminance position refers to the space coordinates and corresponding illuminance value of the measurement point in the task area illuminance measurement process, which is input into the matrix structure according to the space mapping sequence, and the illumination distribution of the differentiated position and the difference range are analyzed; The coverable LED lighting node refers to the illumination coverage range data of the LED lamp node in the illuminance difference area processing process, the difference area coordinates are compared with the node illumination range in space, the coverage ratio is calculated, and the node capable of lighting is selected; The LED lighting node driving control duty cycle refers to the proportion value of the high level time of each LED node PWM driving signal to the whole cycle in the brightness adjustment process, which is compared with the power upper limit and current limit value of the node, and the brightness adjustment space is analyzed; The duty cycle adjustment instruction refers to the original duty cycle value after the comparison of duty cycle, power and current, which is modified according to the preset adjustment rule, and the modified value is matched with the node identification and control parameter. 2.The deep learning-based adaptive dimming method for LED lamps according to claim 1, wherein, The task area illuminance distribution information includes sampling point illuminance value, node space position relationship and illuminance distribution matrix, the LED node adjustment list includes node identification information, coverage area data and matching coverage ratio, the node duty cycle adjustment instruction includes node identification information, duty cycle change data, power and current limit parameter, the node phase and duty cycle control information includes node number information, phase delay data and duty cycle control data, and the LED lamp self-matching dimming control instruction includes full node control information, power distribution data and illumination direction parameter. 3.The deep learning-based adaptive dimming method for LED lamps according to claim 1, wherein, The specific steps of S1 are: S101: obtaining the illuminance information and node space coordinate data of the LED lighting node in the task area sampling point, inputting the sampling point illuminance information and PWM control driving power output value into the data table according to the sampling point corresponding relationship, and obtaining the illuminance and power corresponding data; S102: based on the illumination and power corresponding data, calling node space coordinate data, corresponding node position and sampling point position in space, and associating each sampling point to illumination information and power information, to obtain spatially associated illumination information; S103: based on the spatially associated illumination information, according to the spatial mapping order of the nodes and the sampling points, inputting the illumination information under the mapping order to the matrix structure according to the row and column positions, to obtain the task area illumination distribution information. 4.The deep learning-based adaptive dimming method for LED lamps according to claim 1, wherein, The specific steps of S2 are: S201: based on the task area illumination distribution information, calling task area coordinates and target illumination requirement value, comparing the illumination of the sampling points and the target illumination corresponding position, dividing the inconsistent sampling point position range, to obtain the illumination difference position; S202: based on the illumination difference position, calling the coverage range information of the LED lighting node, corresponding matching the illumination difference position and the coverable node, calculating the coverage proportion of the node in the range, screening the coverable node set, to obtain the coverage node set; S203: based on the coverage node set, calling the coverage proportion parameter of the node and the sampling point, distributing the proportion parameter according to the corresponding relationship between the node and the sampling point, to obtain the LED node adjustment list. 5.The deep learning based adaptive dimming method for LED lamps according to claim 1, wherein, The specific steps of S3 are: S301: based on the LED node adjustment list, calling the LED lighting node drive control duty cycle data, corresponding matching the node and the original duty cycle value according to the node identifier, associating the drive channel number and the node identifier, and mapping the relationship, setting the index relationship according to the node sequence, to obtain the duty cycle matching result; S302: based on the duty cycle matching result, comparing the duty cycle matching result with the power upper limit and the current limit value, calculating the current utilization rate of the node, and corresponding the result with the node identifier, to obtain the current utilization result; S303: based on the current utilization result, comparing the current utilization rate with the duty cycle matching data, and replacing the corresponding values, converting into a control instruction format that can be directly issued, to obtain the node duty cycle adjustment instruction. 6.The deep learning based adaptive dimming method for LED lamps according to claim 1, wherein, The specific steps of S4 are: S401: based on the node duty cycle adjustment instruction, extracting the LED lighting node PWM period data, splitting the node period length into sections according to the number, and corresponding with the node number, to obtain the period section information; S402: based on the period section information, assigning phase delay values to the sections in turn, and associating each phase delay with the corresponding node number, to obtain the phase delay assignment result; S403: based on the phase delay assignment result and the node duty cycle adjustment instruction, sequentially associating the phase delay and the duty cycle data according to the node number, outputting the executable instruction structure, to obtain the node phase and duty cycle control information. 7.The deep learning-based adaptive dimming method for LED lamps according to claim 1, wherein, The specific steps of S5 are: S501: based on the node phase and duty cycle control information, calling the real-time power upper limit data of the LED lighting node, associating the power upper limit with the corresponding node number, and corresponding with the control information, to obtain the power matching result; S502: based on the power matching result, according to the illumination direction data of the LED lighting node, sequentially corresponding the direction data with the node number in the power matching result, to obtain the direction power corresponding result; S503: Based on the direction power corresponding result, call the node phase and duty cycle control information, correspond the node phase, duty cycle, power and direction data in turn, convert into executable full node control content, get LED lamp self-matching dimming control instruction.

8. A deep learning based LED lamp adaptive dimming system, characterized in that, The system is used to realize the deep learning-based LED lamp adaptive dimming method of any one of claims 1-7, and the system comprises: The illumination data acquisition module acquires the illumination information and node spatial coordinate data of the LED lighting node at the sampling points in the task area, processes the illumination information and PWM control driving power output value according to the sampling point correspondence relationship, and forms a distribution matrix through the spatial mapping sequence of the nodes and the sampling points to obtain the task area illumination distribution information. The task area matching module, based on the task area illumination distribution information, calls the task area coordinate and target illumination demand value, compares the positions of the sampling points, matches the insufficient illumination sampling points with the coverable LED lighting nodes, and combines them according to the node and sampling point coverage ratio to obtain an LED node adjustment list. The duty cycle adjustment module, based on the LED node adjustment list, calls the LED lighting node driving control duty cycle data, matches the node and the original duty cycle value according to the node identification, compares the matching result with the node power upper limit and the current limit value to obtain a node duty cycle adjustment instruction. The phase control module, based on the node duty cycle adjustment instruction, calls the LED lighting node PWM period data, distributes the phase delay value after the node period length is divided by the number, and combines the phase delay value and the corresponding duty cycle adjustment instruction according to the node number sequence to obtain the node phase and duty cycle control information. The dimming instruction generation module, based on the node phase and duty cycle control information, calls the LED lighting node real-time power upper limit and illumination direction data, corresponds the control information with the power and direction data, converts into executable full node control content, and obtains the LED lamp self-matching dimming control instruction.

Images