Camera-based laser impact point detection device and method

Abstract

To provide a camera-based laser impact point detection device and method that can suppress false detections and improve the reliability of detection results. [Solution] This invention relates to a camera-based laser impact point detection device and method. It extracts a region of change by the difference between an image frame generated by capturing a target and a reference background image, and removes noise and reflected light by contour detection and area-based filtering. Furthermore, it identifies the actual laser impact point and calculates the center coordinates by performing hue-saturation-brightness (HSV) color space analysis and core analysis based on pixel distribution. In addition, it stores frames and metadata before and after the detection point, and uses these to analyze the consistency of brightness changes and positions, thereby preventing false detections and improving the reliability of the detection results.

Description

Technical Field

[0001] The present invention relates to a laser impact point detection technology based on image processing, and more particularly, to a technology for detecting and verifying a laser impact point irradiated onto a target using a camera. This can be applied to the fields of image analysis, image-based measurement, and optical detection.

Background Art

[0002] In recent years, technologies for detecting laser impact points on a target using a camera have been widely used in the fields of shooting training, sports analysis, and interactive content. Such a technology has advantages in terms of cost reduction and ease of installation, in that it can calculate the impact position through image analysis without using a separate sensor.

[0003] As an example, US Patent Application Publication No. 2023 / 0332871 discloses a technology that detects a changed region on a target using the difference between an image captured by a camera and a reference image, and determines the impact point based thereon. The above technology presents a structure for detecting impact point candidates through the setting of a Region of Interest (ROI) and an image processing algorithm, and provides a basic framework for image-based impact point detection.

[0004] However, such conventional technologies mainly rely on the difference between images or changes in brightness to detect impact points, so there is a problem that it is difficult to distinguish regions having luminance characteristics similar to actual laser signals due to reflected light generated on the target surface or changes in ambient illumination. In particular, in the case of metallic surfaces or glossy materials, reflected light appears strongly and can be confused with laser signals, which leads to frequent false detections, resulting in the limitation that detection accuracy and reliability are reduced.

[0005] Furthermore, conventional technologies lack sufficient mechanisms to verify the accuracy of the results after the point of impact has been detected. This makes it difficult to retrospectively judge or correct detection errors. In other words, methods that derive results based on a single frame or instantaneous image information do not consider temporal continuity or consistency of positional changes, thus limiting the objective reliability of the detection results.

[0006] Therefore, there is a need to develop a technology that can effectively suppress false detections due to reflected light and noise, while verifying the detection results of the laser impact point based on temporal and spatial information. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] U.S. Patent Application Publication No. 2023 / 0332871 (Publication Date: October 19, 2023) [Overview of the project] [Problems that the invention aims to solve]

[0008] This invention was made to solve the above-mentioned problems, and aims to solve the problem of false detections occurring because it is not possible to precisely distinguish regions having brightness characteristics similar to the actual laser signal due to changes in reflected light and ambient illumination generated on the target surface.

[0009] Furthermore, the present invention aims to overcome the limitations of the conventional technology, where it is difficult to objectively determine the reliability of detection results due to a detection method based on a single frame, and to provide a camera-based laser impact point detection device and method that can verify the detection results of laser impact points based on temporal and spatial information.

[0010] The technical problems that the present invention aims to solve are not limited to those mentioned above, and other technical problems not mentioned can be clearly understood by a person with ordinary skill in the art to which the present invention pertains from the following description. [Means for solving the problem]

[0011] To achieve the above objective, an apparatus for detecting the point of laser impact on a target using a camera through image processing, according to one aspect of the present invention, includes: an image generation unit that continuously captures an area including the target and generates image frames; a background image generation unit that generates and stores a reference background image using a plurality of image frames from the image frames input from the image generation unit that are generated in an initial state when the laser is not irradiated; a change detection unit that extracts a change region by performing a difference calculation between the current frame input from the image generation unit and the reference background image on a pixel-by-pixel basis; a contour detection unit that detects contours for a plurality of laser candidate regions included in the change region; and a function that calculates the area of ​​the contours detected by the contour detection unit, removes contours below a minimum threshold as contours corresponding to noise, and removes contours exceeding a maximum threshold as contours corresponding to reflected light. The system is characterized by including: a filtering unit that selects laser candidate regions; a color analysis unit that converts the laser candidate regions selected by the filtering unit into a Hue, Saturation, Value (HSV) color space and distinguishes between the laser center region and the reflected light region using the lightness value and saturation value; a core analysis unit that analyzes the pixel distribution inside the laser center region and determines whether or not it is an actual laser impact point based on the presence or absence of pixels with higher brightness values ​​compared to the surrounding area and the spatial density of such pixels; a coordinate calculation unit that calculates the geometric center coordinates of the region determined by the core analysis unit to be an actual laser impact point; and a background update control unit that, for frames in which the laser impact point has been determined by the coordinate calculation unit, controls the system so that such frames are not reflected in the background image update to prevent contamination of the background image.

[0012] Furthermore, the camera-based laser impact point detection device further includes a frame storage unit that stores an event frameset including multiple image frames before the point in time when the laser impact point is detected by the core analysis unit, an image frame at that point in time, and multiple image frames after that point in time. The frame storage unit is characterized in that it stores the event frameset in association with the center coordinates of the laser impact point calculated by the coordinate calculation unit.

[0013] Furthermore, the event frameset is characterized by including the two frames prior to the detection of the laser impact point, the one frame at that point, and the two frames following it.

[0014] Furthermore, the frame storage unit is characterized by storing metadata, including the center coordinates of the laser impact point, the score, and the detection time, in association with the event frameset.

[0015] Furthermore, the laser impact point detection device based on the camera is characterized by further comprising a verification unit that uses the event frameset and metadata stored in the frame storage unit to analyze the consistency of brightness changes and positions between preceding and succeeding frames, with reference to the time when the laser impact point was detected, and to determine whether or not a false detection has occurred.

[0016] On the other hand, a method performed in an apparatus for detecting the point of laser impact on a target using a camera by image processing, according to another aspect of the present invention, includes the steps of: an image generation unit continuously capturing images of a shooting area including the target to generate image frames; a background image generation unit generating and storing a reference background image using a plurality of image frames from the image frames input from the image generation unit that are generated in an initial state when the laser is not irradiated; a change detection unit extracting a change region by performing a difference calculation between the current frame input from the image generation unit and the reference background image on a pixel-by-pixel basis; a contour detection unit detecting contours for a plurality of laser candidate regions included in the change region; and a filtering unit calculating the area of ​​the contours, removing contours below a minimum threshold as contours corresponding to noise, and removing contours exceeding a maximum threshold as reflected light The invention is characterized by comprising: a step of selecting a laser candidate region by removing it as a corresponding contour; a step of a color analysis unit converting the laser candidate region selected by the filtering unit into a hue-saturation-lightness (HSV) color space and distinguishing the laser center region and the reflected light region using lightness and saturation values; a step of a core analysis unit analyzing the pixel distribution inside the laser center region and determining whether or not it is an actual laser impact point based on the presence or absence of pixels with higher brightness values ​​compared to the surrounding area and the spatial density of such pixels; a step of a coordinate calculation unit calculating the geometric center coordinates of the region determined by the core analysis unit to be an actual laser impact point; and a step of a background update control unit controlling so that the frame in which the laser impact point has been determined by the coordinate calculation unit is not reflected in the background image update, thereby preventing contamination of the background image.

[0017] Furthermore, the laser impact point detection method based on the camera is characterized by comprising: the step of the frame storage unit storing an event frame set including a plurality of image frames before the point in time when the laser impact point is detected by the core analysis unit, an image frame at that point in time, and a plurality of image frames after that point in time; and the step of the frame storage unit storing the event frame set in association with the center coordinates of the laser impact point calculated by the coordinate calculation unit.

[0018] Furthermore, the step of the frame storage unit storing the event frameset in association with the center coordinates of the laser impact point calculated by the coordinate calculation unit is characterized in that the frame storage unit further includes the step of the frame storage unit storing metadata including the center coordinates of the laser impact point, score, and detection time in association with the event frameset.

[0019] Furthermore, the laser impact point detection method based on the camera is characterized by further including a step in which the verification unit uses the event frameset and metadata stored in the frame storage unit to analyze the consistency of brightness changes and positions between preceding and succeeding frames, with reference to the time when the laser impact point was detected, and thereby determine whether or not a false detection has occurred. [Effects of the Invention]

[0020] The present invention achieves the following effects through the above-described solution.

[0021] Firstly, according to the present invention, after detecting a change region based on the difference with the background image, noise and reflected light are removed through contour-based filtering, and by combining HSV color space analysis and core analysis based on pixel distribution, precise discrimination that reflects the characteristics of the laser impact point is possible. This effectively suppresses false detections due to reflected light and environmental changes.

[0022] Secondly, by storing an event frameset including frames before and after the detection point of the laser impact, along with metadata corresponding to that frame, and using this to analyze the consistency of brightness changes and position, the reliability of the detection results can be objectively determined. This enables post-event verification of the detection results, improving the reliability and fairness of the system.

[0023] Third, by excluding the frame in which the laser landing point is detected from the update of the background image, contamination of the background image can be prevented, and stable detection performance can be maintained even in a repetitive detection environment.

[0024] The effects of the present invention are not limited to the effects described above, and should be understood to include all effects that can be inferred from the configuration of the invention described in the detailed description of the present invention or the claims. [Brief Description of the Drawings]

[0025] [Figure 1] It is a schematic diagram showing the overall configuration of a camera-based laser landing point detection system according to the present invention. [Figure 2] It is a functional configuration block diagram of a camera-based laser landing point detection apparatus 200 according to the present invention. [Figure 3] It is a screen showing an example of the detection result of a camera-based laser landing point according to the present invention. [Figure 4] It is a flowchart showing the processes of detecting a change region and screening laser candidate regions in the camera-based laser landing point detection method according to the present invention. [Figure 5] It is a flowchart showing the processes of precise discrimination and verification of laser landing points in the camera-based laser landing point detection method according to the present invention. [Figure 6] It is a conceptual diagram showing a structure in which an event frame set is configured with reference to the detection time point of a laser landing point according to the present invention, and metadata is associated and stored. [Mode for Carrying Out the Invention]

[0026] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, so that those skilled in the art in which the present invention pertains can easily implement it. The present invention can be implemented in various different forms and is not limited to the embodiments described below. In the drawings, parts that are not relevant to the description have been omitted in order to clearly illustrate the present invention, and the same or similar components are denoted by the same reference numerals throughout this specification.

[0027] The words and terms used herein and in the claims shall not be interpreted to be limited to their ordinary or dictionary meanings, but rather to be interpreted in a way that is consistent with the technical idea of ​​the present invention, in accordance with the principle that inventors may define terms and concepts in order to best describe their invention.

[0028] Therefore, the embodiments described herein and the configurations shown in the drawings represent one preferred embodiment of the present invention and do not represent the entire technical concept of the present invention. As a result, there may be various equivalents and variations of these configurations that can be substituted at the time of filing of the present invention.

[0029] In this specification, terms such as “includes” or “has” are intended to describe the presence of features, numbers, steps, actions, components, parts, or combinations thereof described herein, and should be understood not to preemptively exclude the possibility of the presence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. Furthermore, “connected” to another component includes not only direct connection but also indirect connection, unless otherwise specified.

[0030] In this specification, singular expressions include plural expressions unless the context clearly indicates singular. Similarly, plural expressions include singular expressions unless the context clearly indicates plural. Throughout this specification, when a part contains a component, it does not exclude other components, but rather means that other components may be included, unless otherwise stated.

[0031] Preferred embodiments of the present invention will be described in detail below with reference to the attached drawings.

[0032] Figure 1 is a schematic diagram showing the overall configuration of the camera-based laser impact point detection system according to the present invention.

[0033] Referring to Figure 1, the system includes a laser device 11 that emits a laser, a target 12 that is illuminated by the laser, a camera 13 that photographs the target, and image processing devices 21 and 22 that detect the laser impact point based on the captured image.

[0034] In recent years, technologies that use cameras to detect laser impact points on targets have been utilized in shooting training, sports analysis, and interactive content. Such technologies are advantageous in terms of simplifying system configuration and reducing costs, as they can calculate impact positions through image analysis without using separate physical sensors, and they are also easy to install and operate. The laser device 11 is configured to illuminate the target 12 with a laser in response to user input. The target 12 is the target area illuminated by the laser and may include multiple concentrically divided score areas, providing a basis for calculating scores according to the laser impact position.

[0035] Camera 13 is positioned facing target 12 and continuously captures images including the target area, and transmits the captured image frames to the image processing unit. Camera 13 continuously captures the shooting area including target 12 to generate a time-series of image frames. At this time, camera 13 captures images at a constant frame rate, for example, 60 frames per second, and the generated image frames are transmitted to the processing unit 22.

[0036] Conventionally, a method was used in which an electronic sensor was attached to the target surface to detect the point of impact. However, this method had the drawbacks of increased costs associated with sensor installation and maintenance, and reduced reliability due to physical damage or malfunction of the sensor due to repeated use. In the present invention, the target is photographed in a non-contact manner using a general-purpose camera 13, and the point of laser impact is detected by image analysis. This makes it possible to detect the point of impact without attaching a separate sensor. This simplifies the system configuration, reduces installation and maintenance costs, and improves the reliability and durability of the device by eliminating failure factors associated with physical contact. In particular, the present invention goes beyond simple image capture and enables stable point of impact detection while suppressing the problem of false detection that could not be solved with sensor methods, by combining background subtraction, contour-based filtering, color analysis, and core analysis.

[0037] The image processing device may include a display device 21 and a processing unit 22. The processing unit 22 analyzes the image frame input from the camera 13 to detect the laser impact point and outputs the detection result to the display device 21.

[0038] The processing unit 22 receives image frames input from the camera and performs background image generation, detection of change regions by inter-frame differences, contour extraction, and area-based filtering. It also determines the actual laser impact point through HSV color space analysis and core analysis based on pixel distribution, and calculates the center coordinates. Furthermore, it stores frames and metadata before and after the detection point, analyzes the consistency of brightness changes and positions to verify the absence of false detections, and controls background image updates to maintain stable detection performance.

[0039] Figure 2 is a functional block diagram of the camera-based laser impact point detection device 200 according to the present invention.

[0040] Referring to Figure 2, the apparatus 200 may include an image generation unit 210, a background image generation unit 215, a change detection unit 220, a contour detection unit 225, a filtering unit 230, a color analysis unit 235, a core analysis unit 240, a coordinate calculation unit 245, a background update control unit 250, a frame storage unit 255, a verification unit 260, and a result output unit 265.

[0041] The image generation unit 210 includes a camera positioned toward a shooting area containing a target, and continuously captures the shooting area to generate a time-series of image frames. The image generation unit 210 can generate image frames at a speed of, for example, 30 frames per second or more, preferably 60 frames per second or more, and the generated image frames are provided as processing data for the background image generation unit 215, the change detection unit 220, the frame storage unit 255, and the verification unit 260.

[0042] Thus, the image generation unit 210 acquires changes in the target's state in a non-contact manner using a general-purpose camera without attaching a separate electronic sensor to the target. This makes the device easy to install and reduces detection errors due to physical damage to the target or sensor failure.

[0043] The background image generation unit 215 generates and stores a reference background image using multiple image frames input from the image generation unit 210 in the initial state when the laser is not irradiated. The reference background image is then used as reference data to be compared with the current frame, and reflects the target surface, ambient lighting, and shooting environment before laser irradiation.

[0044] The background image generation unit 215 can mitigate camera noise and instantaneous lighting fluctuations by averaging or accumulating multiple initial frames. As a result, the change detection unit 220 can perform difference calculations based on a stabilized reference background image rather than a single frame, thereby suppressing the detection of unwanted changes other than the laser signal as candidate regions.

[0045] The change detection unit 220 extracts the change region by performing a pixel-by-pixel difference calculation between the current frame input from the image generation unit 210 and the reference background image generated by the background image generation unit 215. The change region may include not only the region corresponding to the local increase in brightness caused by laser irradiation, but also the region corresponding to the reflected light generated on the target surface, or the region corresponding to the change in brightness due to a change in external lighting.

[0046] As a result, the change detection unit 220 does not simply determine a laser impact point based on whether it is a bright area, but rather extracts areas where changes have occurred based on the difference with the background image as candidate areas and transmits them to the subsequent processing stage. This structure comprehensively detects all change elements, including reflected light (light blurring) generated during laser irradiation, thereby providing basic data for precise distinction in the subsequent contour detection unit 225 and filtering unit 230.

[0047] Furthermore, the change detection unit 220 can be linked to the background update control unit 250 to exclude frames detected by the laser from updating the background image, as detection accuracy may decrease if such frames are reflected in the background model. This prevents contamination of the background image and enables stable change detection even in continuous detection environments.

[0048] The contour detection unit 225 analyzes the boundaries of connected pixel groups for the change region extracted by the change detection unit 220 to detect the contour and separates multiple laser candidate regions into object units. Rather than simply processing bright pixels as a collection, the contour detection unit 225 defines spatially connected regions as a single candidate object, thereby providing a basis for structurally distinguishing the laser impact point from reflected light or noise. In particular, when a laser is irradiated, the target surface may simultaneously generate reflected light that diffuses to the periphery along with high brightness at the center, and such reflected light has the characteristic of spreading over a wide area.

[0049] The contour detection unit 225 reflects these brightness distribution characteristics, separates the change region into individual contour units, and transmits this information to the filtering unit 230 so that the area and shape analysis of each contour can be performed. As a result, the subsequent filtering unit 230 can effectively remove excessively small contours (e.g., noise) and excessively large contours (e.g., reflected light) using the contour area information. Furthermore, by precisely structuring the change region, the contour detection unit 225 provides input data that enables the color analysis unit 235 and the core analysis unit 240 to perform pixel distribution analysis on the actual laser center region, thereby enabling accurate identification of the laser impact point.

[0050] The filtering unit 230 calculates the area of ​​each contour detected by the contour detection unit 225, and compares the calculated area with a preset minimum threshold (e.g., 5px) and maximum threshold (e.g., 5000px) to select candidate laser regions. Contours below the minimum threshold are removed because they are likely to be camera noise, fine dust, sensor noise, or image compression noise.

[0051] Conversely, contours exceeding the maximum threshold are likely to be areas caused by reflected light diffused over a wider area than the actual laser impact point or by changes in external illumination, and are therefore removed. In this way, the filtering unit 230 improves processing efficiency and detection stability by temporarily narrowing down candidate areas using the size characteristics of the laser impact point, thereby preventing the color analysis unit 235 and core analysis unit 240 from analyzing unnecessary areas.

[0052] The color analysis unit 235 converts the laser candidate region selected by the filtering unit 230 into the HSV color space and distinguishes between the laser center region and the reflected light region using the lightness value (V) and chroma value (S). Because the HSV color space separates and represents hue, chroma, and lightness, it is more advantageous for distinguishing between brightness changes and color bleeding than when using only RGB values.

[0053] The laser center region generally has a higher brightness value compared to the surrounding area, but may exhibit relatively low saturation or near-white characteristics at the center. On the other hand, reflected light generated at the target surface may diffuse to the periphery or be accompanied by color bleeding. Therefore, by using a combination of brightness and saturation values, the laser center region and the reflected light region can be distinguished more precisely. The analysis results from the color analysis unit 235 are transmitted to the core analysis unit 240 and used not only for simple brightness determination but also for final discrimination based on pixel distribution.

[0054] The core analysis unit 240 analyzes the pixel distribution within the region identified as the laser center region by the color analysis unit 235 to determine whether or not it is the actual laser impact point. Specifically, the core analysis unit 240 analyzes whether or not pixels with higher brightness values ​​than the surrounding area exist within the region, and how spatially densely those pixels are located.

[0055] At actual laser impact points, a relatively small region with concentrated luminance values ​​is generally formed in the center, while reflected light tends to spread widely or exhibit an irregular distribution. Therefore, the core analysis unit 240 does not simply determine a laser based on the brightness of the region alone, but also considers the spatial density and centrality of the luminance values ​​to identify the actual laser impact point. This core analysis, performed as the final discrimination step after change detection, contour detection, area filtering, and HSV color analysis, is a core component that substantially realizes the false detection suppression effect of the present invention.

[0056] The coordinate calculation unit 245 calculates the geometric center coordinates of the region determined by the core analysis unit 240 to be the actual laser impact point. The center coordinates can be calculated using the position of the pixels constituting the laser impact point region, the center point of the contour, or the luminance centroid.

[0057] The calculated center coordinates can be used for score calculation in correspondence with the target score area, and can also be used as metadata stored in association with the event frameset by the frame storage unit 255. In this way, the coordinate calculation unit 245 not only provides a position output function, but also provides reference data for subsequent verification and result display.

[0058] The background update control unit 250 controls whether or not to update the reference background image based on the determination result of the core analysis unit 240 or the coordinate calculation unit 245. Specifically, it excludes frames in which the laser impact point has been determined so that they are not reflected in the update of the reference background image, and controls the background image to be updated using only frames in which no laser impact point has been detected.

[0059] This type of control is intended to prevent the laser signal from being included in the reference background image. If a frame containing the laser is reflected in the background image, the laser signal may be reduced or distorted in subsequent difference calculations, potentially degrading detection performance. Therefore, the background update control unit 250 maintains the stability of the reference background image in long-term operating environments and prevents a decrease in detection performance even during repetitive laser detection processes.

[0060] The frame storage unit 255 stores an event frameset that includes multiple image frames before the point in time when the laser impact point is detected by the core analysis unit 240, an image frame at that point in time, and multiple image frames after that point in time. For example, as shown in Figure 6, the frame storage unit 255 can store a total of 5 frames, including 2 frames before the detection point, 1 frame at the detection point, and 2 frames after the detection point.

[0061] Furthermore, the frame storage unit 255 stores the event frameset in association with the center coordinates of the laser impact point calculated by the coordinate calculation unit 245. In some cases, metadata including the score and detection time may also be stored along with the event frameset.

[0062] The frame storage unit 255 is not merely a recording means, but also provides temporal and spatial basis data for the verification unit 260 to determine whether or not a false detection has occurred. As a result, the present invention can retrospectively verify the detection results of the laser impact point, thereby ensuring fairness and reliability in score calculation.

[0063] The verification unit 260 uses the event frameset and metadata stored in the frame storage unit 255 to determine whether or not a false detection of the laser impact point has occurred. Specifically, the verification unit 260 analyzes the consistency of brightness changes and positions between frames before and after the time when the laser impact point was detected.

[0064] In the case of an actual laser impact point, a clear change in brightness occurs around the detection point, and the detection position tends to maintain a constant spatial relationship between frames. On the other hand, reflected light and transient noise may have irregular brightness changes or inconsistent positions. The verification unit 260 uses these differences to determine whether the detection result corresponds to an actual laser impact point or is a false detection.

[0065] The verification unit 260 not merely outputs the detection results of the present invention, but also retrospectively determines the reliability of the detection results using the stored event frames and metadata, thereby providing technical features that distinguish it from conventional single-frame-based detection methods.

[0066] The result output unit 265 outputs to the user the center coordinates of the laser impact point calculated by the coordinate calculation unit 245, the score, the detection time, and the false detection judgment result of the verification unit 260. The result output unit 265 can be implemented as a display device, monitor, tablet, mobile terminal, or a separate user interface device.

[0067] The result output unit 265 can display the detected laser impact point on the target image, and also display the score and verification results. This allows the user to confirm not only the laser impact point but also whether the detection result is reliable.

[0068] Figure 3 shows an example of the detection result of the laser impact point based on the camera according to the present invention.

[0069] Referring to Figure 3, the laser impact point is displayed on the target image, appearing as a high-brightness region concentrated in the center. The impact point is a region detected by the color analysis unit 235 and the core analysis unit 240, and is characterized by a dense distribution of pixels with high brightness values ​​in the center.

[0070] Furthermore, a diffuse region generated by reflected light may appear around the laser impact point, and this diffuse region has an irregular and broad distribution that distinguishes it from the actual laser center region. In this invention, such reflected light regions are removed by stepwise processing by the filtering unit 230, the color analysis unit 235, and the core analysis unit 240, and finally only the laser impact points that appear densely concentrated in the center are identified.

[0071] The lower end of Figure 3 may display the center coordinates (X,Y) of the laser impact point, the detection time (Time), distance information (Dist), and score (Score). The coordinate information is calculated by the coordinate calculation unit 245, and the score is determined according to the positional relationship with the score area of ​​the target.

[0072] In particular, the present invention stores an event frameset including frames before and after the point in time when the laser impact point is detected, and uses this to analyze the consistency of brightness changes and position, thereby verifying whether or not the impact point was caused by an actual laser. As a result, it can be confirmed that the impact point displayed as shown in Figure 3 is not merely a detection result, but a highly reliable result that has undergone a verification process.

[0073] Figure 4 is a flowchart showing the process of detecting a change region and selecting a candidate laser region based on its contour, as part of the camera-based laser impact point detection method according to the present invention.

[0074] Referring to Figure 4, first, the image generation unit continuously captures images of the shooting area including the target (S411). Then, it generates multiple image frames in the initial state before the laser is irradiated (S412), and uses these multiple initial frames to generate and store a reference background image (S413). This reference background image reflects stable background information before laser irradiation and is used as reference data for subsequent change detection.

[0075] Subsequently, when the current frame is input (S414), a difference calculation between the current frame and the reference background image is performed on a pixel-by-pixel basis (S415). This extracts a region of change based on the difference in brightness or color between the two images (S416). At this time, the region of change may include not only the region corresponding to the laser impact point, but also the region corresponding to the reflected light generated on the target surface, the region caused by changes in illumination, and the region caused by image noise.

[0076] Contour detection is performed on the extracted change regions, and contours are identified for each spatially connected group of pixels (S417). These contours are not simply collections of bright pixels, but are defined as candidate regions on an object-by-object basis and are used as fundamental data for structurally distinguishing the laser impact point from reflected light or noise.

[0077] Next, the area of ​​each contour is calculated (S418), and it is determined whether the area is less than a preset minimum threshold (S419). Contours below the minimum threshold are likely to be camera noise or minute brightness changes, and are therefore considered noise and removed. For example, very small pixel areas (e.g., a few pixels or less) are unlikely to be considered actual laser impact points and are therefore targeted for removal.

[0078] Furthermore, it is determined whether the area of ​​each contour exceeds the maximum threshold (S420). If it exceeds the maximum threshold, it is determined to be a diffuse region formed by reflected light or light diffusion generated on the target surface and is removed. Generally, laser impact points are limited to relatively small areas, while reflected light spreads over a wide area. Therefore, these two regions can be effectively distinguished through area-based filtering.

[0079] By performing double filtering using the minimum and maximum thresholds described above, only candidate regions that are highly likely to correspond to the actual laser impact point are selected (S421). In other words, Figure 4 shows the process of detecting the change region by the difference with the background image, and then removing noise and reflected light by combining contour detection and area-based analysis to gradually narrow down the candidate laser regions.

[0080] Furthermore, in this invention, if a frame in which a laser impact point is detected is reflected in the background image update, subsequent detection accuracy may decrease. Therefore, it is possible to control the system to intentionally exclude such frames from the background model update. This prevents contamination of the background image and enables stable change detection and candidate region selection even in repetitive detection environments.

[0081] Figure 5 is a flowchart showing the process of precise identification and verification of the laser impact point in the camera-based laser impact point detection method according to the present invention.

[0082] Referring to Figure 5, the laser candidate regions selected in Figure 4 are converted to the HSV color space, and their brightness and chroma values ​​are analyzed (S511). This distinguishes between the laser center region and the reflected light region (S512).

[0083] Subsequently, the pixel distribution and spatial density within the laser center region are analyzed (S513). This analysis determines whether or not the region is the actual laser impact point (S514). If it is determined to be the actual laser impact point, the geometric center coordinates of the region are calculated (S515).

[0084] Based on the calculated coordinates, an event frameset including frames before and after the time the laser impact point was detected is stored (S516). Along with the event frameset, metadata including coordinates, score, and detection time is also stored (S517).

[0085] Subsequently, the stored event frameset and metadata are used to analyze the consistency of brightness changes and positions between preceding and succeeding frames, thereby determining whether or not a false detection has occurred (S518). Finally, the coordinates of the verified laser impact point, the score, and the verification results are output to the user (S519).

[0086] In other words, Figure 5 shows a series of processes that identify the actual laser impact point through precise discrimination based on color analysis and pixel distribution, and ensure the reliability of the detection results through the storage and verification process of event frames.

[0087] Figure 6 is a conceptual diagram showing a structure according to the present invention that constructs an event frameset based on the detection time of the laser impact point and stores associated metadata.

[0088] Referring to Figure 6, the time when the laser impact point is detected is set as the reference time T, and multiple image frames, including the previous frame (T-2, T-1) and the subsequent frame (T+1, T+2), are configured as an event frameset centered around this time.

[0089] Specifically, the event frameset includes T-2 and T-1 frames showing the state before detection, a T frame at the time the laser impact point was actually detected, and T+1 and T+2 frames showing the changes after detection. Such a frame configuration is used as a data set for analyzing the temporal change characteristics associated with laser irradiation.

[0090] Furthermore, the event frameset stores the center coordinates of the laser impact point calculated by the coordinate calculation unit, score information corresponding to the target score, and detection time (Time) as associated metadata.

[0091] The event frameset and metadata configured in this way are utilized by the verification unit, which analyzes the consistency of brightness changes and laser impact point positions between preceding and succeeding frames included in the event frameset to determine whether the detection result is due to an actual laser or a false detection due to reflected light or noise.

[0092] In particular, in the case of actual laser impact points, a consistent brightness change appears at a specific location centered on the detection time, while reflected light or noise exhibits temporally irregular or spatially inconsistent distributions. Therefore, analysis based on event frames, as shown in Figure 6, can improve the reliability of the detection results.

[0093] Therefore, the structure shown in Figure 6 represents a core feature of the present invention that overcomes the limitations of detection based on a single frame and enables verification of the laser impact point detection result based on temporal and spatial information.

[0094] While embodiments of the present invention have been described, the technical concept of the present invention is not limited to the embodiments presented herein. Those skilled in the art who understand the technical concept of the present invention can easily conceive of other embodiments by adding, changing, or deleting components within the scope of that technical concept, and these should be understood to be included within the scope of the present invention. [Explanation of symbols]

[0095] 11 Laser device 12 targets 13 Cameras 21 Display device 22 Arithmetic Processing Unit 200 Camera-based laser impact point detection device 210 Image generation unit 215 Background Image Generation Unit 220 Change detection unit 225 Contour detection unit 230 Filtering section 235 Color analysis section 240 Core Analysis Unit 245 Coordinate Calculation Unit 250 Background Update Control Unit 255 Frame storage unit 260 Verification Department 265 Result Output Section

Claims

1. A device that uses a camera to detect the point of impact of a laser projectile directed at a target by image processing, An image generation unit that generates an image frame containing the target, A background image generation unit generates and stores a reference background image using multiple image frames from the image generation unit input to the image generation unit, which are generated in an initial state when the laser is not irradiated. A change detection unit performs a pixel-by-pixel difference calculation between the current frame input from the image generation unit and the reference background image to extract the change region. A contour detection unit that detects contours for multiple laser candidate regions included in the aforementioned change region, A filtering unit calculates the area of ​​the contours detected by the contour detection unit, removes contours below a minimum threshold as noise-corresponding contours, and removes contours exceeding a maximum threshold as reflected light-corresponding contours, thereby selecting candidate laser regions. A color analysis unit that converts the laser candidate region selected by the filtering unit into a hue-saturation-value (HSV) color space and distinguishes the laser center region and the reflected light region using the value and saturation value, A core analysis unit analyzes the pixel distribution within the central region of the laser and determines whether or not it is an actual laser impact point based on the presence or absence of pixels with higher brightness values ​​compared to the surrounding area and the spatial density of such pixels. A coordinate calculation unit calculates the geometric center coordinates of the region determined by the core analysis unit to be the actual laser impact point, For frames in which the laser impact point has been determined by the coordinate calculation unit, a background update control unit controls the frame so that it is not reflected in the background image update, thereby preventing contamination of the background image. Features including, A camera-based laser impact point detection device.

2. The laser impact point detection device based on the aforementioned camera is The system further includes a frame storage unit that stores an event frameset including multiple image frames before the point in time when the laser impact point is detected by the core analysis unit, an image frame at that point in time, and multiple image frames after that point in time, with the point in time being the reference point. The frame storage unit is characterized by storing the event frameset in association with the center coordinates of the laser impact point calculated by the coordinate calculation unit. A laser impact point detection device based on a camera as described in claim 1.

3. The event frameset is characterized by including the two frames prior to the detection of the laser impact point, the one frame at that point, and the two frames following it, with reference to the detection point of the laser impact point. A laser impact point detection device based on a camera as described in claim 2.

4. The frame storage unit is, The system is characterized by storing metadata, including the center coordinates of the laser impact point, the score, and the detection time, in association with the event frameset. A laser impact point detection device based on a camera as described in claim 2.

5. The laser impact point detection device based on the aforementioned camera is The invention further includes a verification unit that uses the event frameset and metadata stored in the frame storage unit to analyze the consistency of brightness changes and positions between frames before and after the time when the laser impact point was detected, and to determine whether or not a false detection occurred. A laser impact point detection device based on a camera as described in claim 4.

6. A method performed in a device that uses a camera to detect the point of impact of a laser beam irradiated onto a target by image processing, The image generation unit generates image frames by continuously capturing images of the shooting area including the target, The background image generation unit generates and stores a reference background image using a plurality of image frames from the image frames input from the image generation unit that were generated in an initial state when the laser was not irradiated. The change detection unit performs a pixel-by-pixel difference calculation between the current frame input from the image generation unit and the reference background image to extract the changed region. The contour detection unit performs the step of detecting contours for a plurality of laser candidate regions included in the change region, The filtering unit performs the following steps to select a laser candidate region: it calculates the area of ​​the contour, removes contours below a minimum threshold as contours corresponding to noise, and removes contours exceeding a maximum threshold as contours corresponding to reflected light; The color analysis unit converts the laser candidate region selected by the filtering unit into a hue-saturation-value (HSV) color space, and distinguishes between the laser center region and the reflected light region using the value and saturation value. The core analysis unit analyzes the pixel distribution within the laser center region and determines whether or not it is an actual laser impact point based on the presence or absence of pixels with higher brightness values ​​compared to the surrounding area and the spatial density of such pixels. The coordinate calculation unit calculates the geometric center coordinates of the region that the core analysis unit has determined to be the actual laser impact point, The background update control unit controls the background so that frames in which the laser impact point has been determined by the coordinate calculation unit are not reflected in the background image update, thereby preventing contamination of the background image. Features including, A camera-based laser impact point detection method.

7. The laser impact point detection method based on the aforementioned camera is: The frame storage unit stores an event frameset including multiple image frames before the point in time when the laser impact point is detected by the core analysis unit, an image frame at that point in time, and multiple image frames after that point in time, with the point in time being the reference point. The frame storage unit further includes the step of storing the event frameset in association with the center coordinates of the laser impact point calculated by the coordinate calculation unit, A laser impact point detection method based on a camera as described in claim 6.

8. The step of the frame storage unit storing the event frameset in association with the center coordinates of the laser impact point calculated by the coordinate calculation unit is as follows: The frame storage unit further includes the step of storing metadata, including the center coordinates of the laser impact point, the score, and the detection time, in association with the event frameset. A laser impact point detection method based on a camera as described in claim 7.

9. The laser impact point detection method based on the aforementioned camera is: The verification unit further includes the step of using the event frameset and metadata stored in the frame storage unit to analyze the consistency of brightness changes and positions between frames before and after the time when the laser impact point was detected, and thereby determining whether or not a false detection occurred. A laser impact point detection method based on a camera as described in claim 8.

Images