An urban building light belt anomaly identification method based on unmanned aerial vehicle vision

Abstract

The present application relates to the technical field of unmanned aerial vehicle inspection and urban lighting operation and maintenance, in particular to a city building light belt anomaly identification method based on unmanned aerial vehicle vision, comprising the following steps: unmanned aerial vehicle image acquisition, image preprocessing, light emitting area identification, light belt segmentation identification, light belt structure template construction, template matching and calibration, and light belt loss determination. The city building light belt anomaly identification method based on unmanned aerial vehicle vision can automatically identify the light belt loss area; the structure template is introduced to reduce false detection and missed detection; it is suitable for various building facades and light belt arrangement forms; it is suitable for large-scale urban night scene lighting inspection; and the artificial inspection cost and safety risk are significantly reduced.

Description

Technical Field

[0001] This invention relates to the field of drone inspection and urban lighting maintenance technology, specifically a method for identifying abnormalities in urban building light strips based on drone vision. Background Technology

[0002] With the widespread adoption of urban nightscape lighting projects, numerous building facades are equipped with LED light strips for outlining, facade decoration, and nighttime display. However, during long-term operation, factors such as aging and breakage of the light strips, damage to the power supply or control circuits, and external damage (wind, rain, construction impacts) often cause some light strips to fail or even entire areas to malfunction, severely impacting the overall nightscape effect.

[0003] The current mainstream inspection methods mainly include: manual visual inspection, manual high-altitude maintenance, and manual viewing of images or videos based on drones. Public literature already contains: schemes for inspecting urban facilities using drones, state recognition schemes for streetlights or individual lighting equipment, and simple luminous target detection methods based on brightness thresholds. However, the above schemes are mostly for point light sources or regular targets, and are difficult to adapt to the long, continuous, and diverse lighting structures such as building light strips.

[0004] Existing technologies have at least the following shortcomings: they cannot accurately identify areas with missing light strips; relying solely on brightness detection makes it difficult to distinguish between normal unlit areas and areas with damaged light strips; they have poor adaptability to complex building facades; buildings have vertical, horizontal, curved, and multi-proportion light strip arrangements; light pollution, reflection, and window interference are serious problems; they lack prior structural knowledge and overall consistency judgment; they cannot identify areas where "light strips should be present but are not emitting light"; they cannot be verified by combining the arrangement patterns of similar light strips; they are difficult to automate and scale up operation and maintenance; they still require secondary manual confirmation, resulting in low efficiency and high cost.

[0005] To address the aforementioned issues, we propose an improvement: a method for identifying anomalies in urban building light strips based on UAV vision. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a method for identifying anomalies in urban building light strips based on UAV vision, comprising the following steps:

[0008] S1. Drone Image Acquisition: Using a drone equipped with an image acquisition device, the target building is photographed at night according to a preset flight path to obtain image data of the building's facade, including the light strips.

[0009] S2. Image preprocessing: processing the acquired raw images Preprocessing is performed, including converting the image from RGB space to the luminance channel. And normalize the brightness channel: ;

[0010] S3. Illumination Region Recognition: Based on the preprocessed brightness image, the luminous region is identified.

[0011] (1) Initial luminous region segmentation using a luminance threshold (T_l):

[0012] ;

[0013] (2) For binary images Perform morphological dilation operations: ;

[0014] S4. LED Strip Segmentation and Recognition: For each connected region (R_i) in the candidate set of luminous regions, calculate its geometric features:

[0015] (1) Region length With width ;

[0016] (2) The aspect ratio characteristic is defined as: ;

[0017] when When this occurs, the area is identified as a candidate area for the light strip;

[0018] (3) Calculate the main direction angle of the region When adjacent regions satisfy: When the spatial distance is less than a preset threshold, the area is merged into the same light strip segment.

[0019] S5. Construction of LED Strip Structure Template: Construct LED strip structure template M, which includes N theoretical LED strips, and its parameters are expressed as follows: ;

[0020] S6. Template Matching and Calibration: This involves matching the actual detected light strips... With template To perform matching, the distance between the center positions of the light strips is calculated: ;

[0021] when At that time, determine the detection light strip With template light strip Matching;

[0022] If template light strip If no corresponding light strip is found in the detection results, it will be marked as an abnormal area.

[0023] S7. LED Strip Missing Detection: Detect missing LED strips in abnormal areas. .

[0024] As a preferred technical solution of the present invention, the image acquisition resolution of the UAV is preferably not less than 1920×1080; the shooting angle is basically perpendicular to the building facade; and it can simultaneously acquire pose and height information for subsequent calibration.

[0025] As a preferred technical solution of the present invention, the and These represent the maximum and minimum values ​​in the luminance channel, respectively.

[0026] In a preferred embodiment of the present invention, S is a structural element used to cover the halo area around the light strip. Connectivity analysis is performed on the expanded area to remove areas smaller than a threshold. The area.

[0027] As a preferred technical solution of the present invention For the first The theoretical length of the LED strip light The spacing between adjacent light strips, This indicates the direction in which the LED strips are arranged.

[0028] As a preferred technical solution of the present invention This refers to the actual length of the light strip detected.

[0029] when When the k-th light strip is determined to be a missing light strip.

[0030] The beneficial effects of this invention are: this method for identifying abnormalities in urban building light strips based on UAV vision can automatically identify areas with missing light strips; it introduces structural templates to reduce false detections and missed detections; it is adaptable to various building facades and light strip arrangement forms; it is suitable for large-scale urban night scene lighting inspections; and it significantly reduces the cost and safety risks of manual inspections. Attached Figure Description

[0031] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0032] Figure 1 This is an image acquisition and preprocessing effect diagram of an abnormal identification method for urban building light strips based on UAV vision according to the present invention;

[0033] Figure 2 This is a halo recognition effect diagram of the urban building light strip anomaly recognition method based on UAV vision according to the present invention;

[0034] Figure 3This is a diagram illustrating the effect of light strip segmentation and recognition in a method for identifying abnormal light strips in urban buildings based on UAV vision, according to the present invention.

[0035] Figure 4 This is a diagram illustrating the construction effect of a light strip structure template for an anomaly identification method for urban building light strips based on UAV vision, as described in this invention.

[0036] Figure 5 This is a diagram illustrating the missing recognition effect of a method for identifying anomalies in urban building light strips based on UAV vision, according to the present invention. Detailed Implementation

[0037] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0038] Example: Figures 1-5 As shown,

[0039] 1. Implementation environment and hardware conditions

[0040] The following equipment and environmental conditions were used in this embodiment:

[0041] (1) Unmanned aerial vehicles (UAVs):

[0042] Quadrone drone;

[0043] It has stable hovering and path planning capabilities;

[0044] Flight altitude: 20–80 meters;

[0045] (2) Image acquisition device:

[0046] Visible light camera;

[0047] Resolution: 1920×1080;

[0048] Frame rate: ≥30fps;

[0049] It is fixedly mounted below the drone, with the lens facing the exterior of the building;

[0050] (3) Calculation and processing unit:

[0051] It can be installed on the drone itself or on a ground workstation;

[0052] Including processor and memory;

[0053] Used to execute the LED strip missing detection algorithm;

[0054] (4) Communication module:

[0055] Used for data transmission between drones and ground stations;

[0056] 2. Detection Objects and Application Scenarios

[0057] The object of the inspection was a high-rise building that had completed its urban nightscape lighting project. Its facade was equipped with multiple vertical LED light strips, each in its designed state:

[0058] The light strips are aligned in the same direction;

[0059] The light strips are basically the same length;

[0060] The spacing between the light strips is arranged in a fixed ratio.

[0061] 3. Specific Implementation Steps

[0062] Step 1: UAV Inspection and Image Acquisition (corresponding to...) Figure 1 )

[0063] The drone hovers at a certain distance from the building along a preset flight path (allowing a view of the entire building) and collects image data of the building's facade while the lights are on at night.

[0064] Flight speed: hovering

[0065] Shooting angle: Shot directly in front of the building.

[0066] Image saving format: JPEG or RAW

[0067] Step 2: Image preprocessing (corresponding to...) Figure 1 )

[0068] The acquired raw images were processed as follows:

[0069] Convert the image from RGB space to HSV space;

[0070] Extract the luminance channel V;

[0071] Normalize the luminance channel;

[0072] The normalized image was denoised using a 5×5 Gaussian filter.

[0073] Step 3: Identification of Emitting Areas (Halo Recognition) (corresponding to) Figure 2 )

[0074] Identification of luminous regions based on brightness channels:

[0075] Brightness threshold Take 0.65;

[0076] Pixels exceeding the threshold are binarized;

[0077] The binary result is subjected to morphological dilation, with a structuring element radius of 3 pixels.

[0078] By analyzing connected components, regions with an area less than 100 pixels are removed.

[0079] Obtain a candidate set of luminous areas in the building facade.

[0080] Step 4: LED strip segmentation and recognition (corresponding to) Figure 3 )

[0081] For each connected region in the candidate set of luminescent regions:

[0082] The minimum bounding rectangle of the computation region;

[0083] Get the region length and width ;

[0084] When aspect ratio When the time is right, it is determined to be a candidate area for the light strip;

[0085] Adjacent regions with a main direction angle deviation of less than 10° and a spacing of less than 50 pixels are merged.

[0086] Finally, a set of continuous light strip segments is obtained.

[0087] Step 5: Construction of LED strip structure template (corresponding to) Figure 4 )

[0088] Based on architectural design data or the results of the first round of integrity inspection, construct the light strip structure template:

[0089] The template contains 12 vertical light strips;

[0090] The theoretical length of a single light strip is 30 to 35 meters;

[0091] The spacing between the light strips is 1.5 to 1.8 meters;

[0092] The LED strips are arranged vertically.

[0093] Step 6: Template Matching and Spatial Calibration

[0094] Spatial matching is performed between the actual detected light strip area and the template:

[0095] Calculate the distance between the center point of the detection light strip and the center point of the template light strip;

[0096] A match is considered successful when the center distance is less than 0.5 meters.

[0097] If a certain light strip in the template does not match any detection results, it is marked as an abnormal light strip.

[0098] Step 7: LED strip missing detection and output (corresponding) Figure 5 )

[0099] Further analysis of the abnormal light strips:

[0100] Calculate the actual detected length of the light strip;

[0101] When the actual length is less than 70% of the template length, the light strip is determined to be a missing light strip;

[0102] Output the spatial location, length, and image annotation results of the missing light strip.

[0103] 4. Implementation Results

[0104] The above implementation method can achieve the following:

[0105] Automatically identify areas with missing lighting strips on building facades;

[0106] Accurately distinguish between "areas without lights" and "areas with damaged light strips";

[0107] This provides accurate location information for subsequent maintenance.

[0108] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for identifying anomalies in urban building light strips based on UAV vision, characterized in that, Includes the following steps: S1. Drone Image Acquisition: Using a drone equipped with an image acquisition device, the target building is photographed at night according to a preset flight path to obtain image data of the building's facade, including the light strips. S2. Image preprocessing: processing the acquired raw images Preprocessing is performed, including converting the image from RGB space to the luminance channel. And normalize the brightness channel: ; S3. Illumination Region Recognition: Based on the preprocessed brightness image, the luminous region is identified. (1) Initial luminous region segmentation using a luminance threshold (T_l): ； (2) For binary images Perform morphological dilation operations: ; S4. LED Strip Segmentation and Recognition: For each connected region (R_i) in the candidate set of luminous regions, calculate its geometric features: (1) Region length With width ; (2) The aspect ratio characteristic is defined as: ; when When this occurs, the area is identified as a candidate area for the light strip; (3) Calculate the main direction angle of the region When adjacent regions satisfy: When the spatial distance is less than a preset threshold, the areas will be merged into the same light strip segment. S5. Construction of LED Strip Structure Template: Construct LED strip structure template M, which includes N theoretical LED strips, and its parameters are expressed as follows: ; S6. Template Matching and Calibration: This involves matching the actual detected light strips... With template To perform matching, the distance between the center positions of the light strips is calculated: ; when At that time, determine the detection light strip With template light strip Matching; If template light strip If no corresponding light strip is found in the detection results, it will be marked as an abnormal area. S7. LED Strip Missing Detection: Detect missing LED strips in abnormal areas. .

2. The method for identifying anomalies in urban building light strips based on UAV vision according to claim 1, characterized in that, The image acquisition resolution of the drone should preferably be no less than 1920×1080; the shooting angle should be basically perpendicular to the building facade; and it can simultaneously acquire pose and altitude information for subsequent calibration.

3. The method for identifying anomalies in urban building light strips based on UAV vision according to claim 1, characterized in that, The and These represent the maximum and minimum values ​​in the luminance channel, respectively.

4. The method for identifying anomalies in urban building light strips based on UAV vision according to claim 1, characterized in that, S is the structuring element used to cover the halo area around the light strip. Connectivity analysis is performed on the expanded region, and areas smaller than a threshold are removed. The area.

5. The method for identifying anomalies in urban building light strips based on UAV vision according to claim 1, characterized in that, For the first The theoretical length of the LED strip, The spacing between adjacent light strips, This indicates the direction in which the LED strips are arranged.

6. The method for identifying anomalies in urban building light strips based on UAV vision according to claim 1, characterized in that, This refers to the actual detected length of the light strip; when When the time is right, the k-th light strip is determined to be a missing light strip.

Images