A device and method for detecting automobile instrument panel indication

Abstract

The present invention relates to the technical field of automobile instrument panel detection, and in particular to a device and method for detecting automobile instrument panel indications, the method comprising using an image acquisition module to capture an image of an automobile instrument panel to obtain an instrument panel image; controlling a mechanical arm module to move to the front of the automobile instrument panel to acquire an image according to the position information of the instrument panel in the instrument panel image; preprocessing the acquired instrument panel image to extract key features of a pointer in the instrument panel image; executing an instrument panel indication detection algorithm according to the extracted key features to identify indication information on the automobile instrument panel; and outputting the identified instrument panel indications. The present invention can solve the technical problems of low efficiency, low accuracy, and difficulty in adapting to complex and variable detection requirements in traditional automobile instrument panel detection methods.

Description

Technical Field

[0001] The invention relates to the technical field of automobile instrument panel detection, and in particular to a device and method for detecting automobile instrument panel indications. Background Art

[0002] With the rapid development of automotive technology, the dashboard screen inside the car has become a key factor in improving the driving experience. They can display various vehicle information such as vehicle speed and engine speed, bringing drivers a richer and more intuitive interactive experience. However, as the number and complexity of screens continue to increase, ensuring their stable operation has become crucial.

[0003] Under the current technical background, the detection of automobile dashboard screens mainly relies on manual inspection or basic automation systems. Manual inspection is not only time-consuming and laborious, but also easily affected by subjective factors, resulting in low detection efficiency and difficulty in ensuring accuracy, making it difficult to meet the actual needs of large-scale production. On the other hand, although simple automation systems can reduce the burden of manpower to a certain extent, in the face of increasingly complex screen structures and diverse fault conditions, their detection accuracy and coverage are often insufficient, and they are prone to missed detection or misjudgment. Summary of the invention

[0004] One purpose of the present invention is to provide a device for detecting the indications of a car dashboard, which can solve the technical problems of low efficiency, low accuracy and difficulty in adapting to complex and changeable detection requirements in traditional car dashboard detection methods.

[0005] Another object of the present invention is to provide a method for detecting the indications on a vehicle instrument panel. To achieve this object, a device for detecting the indications on a vehicle instrument panel as described above is used.

[0006] The present invention adopts the following technical solutions:

[0007] A device for detecting the indication of a vehicle instrument panel, comprising:

[0008] An image acquisition module is used to capture an image of a car dashboard to obtain an image of the dashboard;

[0009] A mechanical arm module, connected to the image acquisition module, and used to move to the front of the vehicle dashboard to acquire an image according to the position information of the vehicle dashboard in the dashboard image taken by the image acquisition module;

[0010] An image processing and analysis module, used for receiving the dashboard image acquired by the image acquisition module, and executing a dashboard indication detection algorithm to identify indication information on the vehicle dashboard;

[0011] The output module is used to output the identified vehicle dashboard indication.

[0012] Preferably, the image acquisition module includes a camera and a camera control unit;

[0013] The camera is used to capture the image of the vehicle dashboard to obtain a dashboard image;

[0014] The camera control unit is used to adjust the shooting parameters of the camera.

[0015] Preferably, the robotic arm module comprises a robotic arm and a robotic arm control unit;

[0016] The mechanical arm is used to move to the front of the automobile dashboard according to the position information;

[0017] The robotic arm control unit is used to accurately control the movement of the robotic arm.

[0018] Preferably, the image processing and analysis module includes a preprocessing unit, a feature extraction unit and an indication recognition unit;

[0019] The preprocessing unit is used to preprocess the collected instrument panel image;

[0020] The feature extraction unit is used to extract key features in the instrument panel image;

[0021] The indication recognition unit is used to recognize the indication information on the vehicle dashboard according to the extracted key features.

[0022] A method for detecting the indication of a vehicle instrument panel, applied to the device for detecting the indication of a vehicle instrument panel as described above, comprises the following steps:

[0023] S1. Use an image acquisition module to capture an image of a car dashboard to obtain a dashboard image;

[0024] S2, according to the position information of the dashboard in the dashboard image, control the robot arm module to move to the front of the car dashboard to collect the image;

[0025] S3, preprocessing the collected dashboard image to extract key features of the pointer in the dashboard image;

[0026] S4, executing the instrument panel indication detection algorithm according to the extracted key features to identify the indication information on the automobile instrument panel;

[0027] S5. Output the recognized instrument panel indication.

[0028] Preferably, in S3, the preprocessing of the collected dashboard image specifically includes the following steps:

[0029] S31. Set the range of hue H, saturation S and brightness V in the dashboard image, and filter by the following conditions:

[0030] H min ≤H≤H max

[0031] S min ≤S≤S max

[0032] V min ≤V≤V max

[0033] Among them, H min , H max , S min , S max 、V min and V max By presetting;

[0034] S32. Use Gaussian blur formula for denoising:

[0035]

[0036] Among them, ω(i,j) represents the Gaussian weight, k represents the window size, and I(x,y) represents the original image;

[0037] S33, adopting a fault tolerance mechanism to appropriately increase or decrease the threshold range of hue, saturation and brightness;

[0038] S34. According to the preset scaling factor, the image pyramid formula is used to perform feature detection to generate a dashboard image of a preset scale:

[0039]

[0040] Among them, s is the scaling factor used to generate images of different scales.

[0041] Preferably, in S3, extracting the key features of the pointer in the dashboard image includes acquiring edge information of the car dashboard using an edge detection algorithm, which specifically includes the following steps:

[0042] S35. Apply the Canny edge detection algorithm to the dashboard image of the preset scale to obtain edge information:

[0043] Set the gradient amplitude of the image G(x,y), let G x and G y is the gradient of the image in the horizontal and vertical directions, and the relationship between the two is:

[0044]

[0045] S36. According to the gradient amplitude G(x,y) of the image, a high threshold high_threshol and a low threshold low_threshold are applied to distinguish edge information:

[0046] When G(x,y)>high_threshol, it is considered a strong edge;

[0047] When low_threshold≤G(x,y)≤high_threshold, it is considered a weak edge;

[0048] When G(x,y)<low_threshold, it is considered as a non-edge.

[0049] Preferably, in S3, extracting the key features of the pointer in the dashboard image includes acquiring the periphery of the car dashboard using a circle detection algorithm, which specifically includes the following steps:

[0050] S37, using the edge information of the image extracted in step S36 as input, using Hough transform to find the circular structure on the dashboard, representing the periphery of the dashboard:

[0051] Define the peripheral equation of the instrument panel as: (xa) 2 +(yb) 2 =r 2

[0052] Among them, (a, b) represents the coordinates of the center of the circle, and r represents the radius of the circle;

[0053] S38. For each edge point (x, y) in the image, consider all circles with different radii r to which it belongs; enumerate different circle center coordinates a, b and radii r, and record the voting results of each group (a, b, r);

[0054] S39, each time the edge points in the image are traversed, the circle accumulator is updated according to the circle parameter space, and the votes of different circle centers and radii are recorded; if multiple edge points vote for the same circle parameter combination, it indicates that the circle exists in the image;

[0055] S310, finding the edge point with the most votes in the circular accumulator, which represents the most likely circular structure in the image, that is, the periphery of the automobile dashboard.

[0056] Preferably, in S3, extracting the key features of the pointer in the dashboard image includes obtaining the pointer of the car dashboard using a line detection algorithm, which specifically includes the following steps:

[0057] S311, using the edge information of the image extracted in step S36 as input, using Hough transform to find the straight line structure on the dashboard, representing the pointer of the dashboard, and using two-dimensional parameters ρ and θ in the parameter space to represent the equation of the straight line:

[0058] ρ=xcos(θ)+ysin(θ)

[0059] Where ρ is the shortest distance from the origin to the line, and θ is the angle between the line and the x-axis;

[0060] S312, for each edge point (x, y) in the image, calculate the parameters ρ and θ of all the straight lines to which it belongs;

[0061] S313, voting the (ρ, θ) of each edge point to the corresponding position in the parameter space, forming a linear accumulator, and recording different (ρ, θ) combinations;

[0062] S314, finding the (ρ, θ) combination with the most votes in the line accumulator, which represents the most likely straight line structure in the image, that is, the pointer of the car dashboard.

[0063] Preferably, in S4, executing the instrument panel indication detection algorithm according to the extracted key features to identify the indication information on the automobile instrument panel specifically includes the following steps:

[0064] S41, the equation of the straight line obtained by line detection: ρ = xcos (θ) + ysin (θ) to obtain the angle θ′ between the pointer and the vertical direction;

[0065] S42, mapping the relationship between the angle θ′ and the scale of the car dashboard, calculating the proportional relationship between the angle and the reading, and obtaining a corresponding mathematical formula.

[0066] One of the above technical solutions has the following beneficial effects:

[0067] 1. High efficiency: The precise positioning of the robotic arm module and the rapid shooting of the image acquisition module greatly improve the efficiency of instrument panel indication detection.

[0068] 2. Accuracy: Using advanced image processing technology and machine learning algorithms, it can accurately extract key features in dashboard images, accurately identify reading information, and reduce the misjudgment rate.

[0069] 3. Automation: The entire testing process is highly automated, reducing manual intervention and labor costs while improving the objectivity and consistency of testing.

[0070] 4. Wide applicability: It can be applied to various types and models of car dashboards, with strong versatility and flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 The present invention is a flow chart of a method for detecting automobile instrument panel indications. DETAILED DESCRIPTION

[0072] The technical solution of the present invention is further described below with reference to the accompanying drawings and through specific implementation methods.

[0073] In the description of the present invention, it is to be understood that the terms “center”, “longitudinal”, “lateral”, “length”, “width”, “thickness”, “up”, “down”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside”, “outside”, “axial”, “radial”, “circumferential”, etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0074] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0075] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0076] A device for detecting the indication of a vehicle instrument panel, comprising:

[0077] An image acquisition module is used to capture an image of a car dashboard to obtain an image of the dashboard;

[0078] A mechanical arm module, connected to the image acquisition module, and used to move to the front of the vehicle dashboard to acquire an image according to the position information of the vehicle dashboard in the dashboard image taken by the image acquisition module;

[0079] An image processing and analysis module, used for receiving the dashboard image acquired by the image acquisition module, and executing a dashboard indication detection algorithm to identify indication information on the vehicle dashboard;

[0080] The output module is used to output the identified vehicle dashboard indication.

[0081] First, the device of the present invention realizes efficient and accurate instrument panel image acquisition. By combining camera positioning and mechanical motion technology, the position of the instrument panel can be accurately determined on the image taken by the camera, and the mechanical arm can be driven to move quickly to the position for image acquisition. This method not only improves the acquisition efficiency, but also ensures the accuracy of the image, laying a solid foundation for subsequent detection work.

[0082] Secondly, the device of the present invention realizes the automation of the detection process. The automatic movement of the robot arm reduces the need for manual operation, making the entire detection process smoother and more efficient. This not only reduces the detection cost, but also reduces the impact of human factors on the detection results, and improves the objectivity and accuracy of the detection.

[0083] Furthermore, the device of the present invention has extremely strong reliability. By collecting dashboard images with a camera and applying advanced dashboard indication detection algorithms, the indication information on the dashboard can be quickly and accurately identified. This provides the driver with timely and accurate vehicle information, effectively improving driving safety and comfort.

[0084] Finally, the device of the present invention has wide applicability. Whether it is a traditional mechanical instrument panel or a modern digital instrument panel, it can perform effective indication detection. This has strong versatility and adaptability, so that it can be widely used in various types and models of cars.

[0085] To further illustrate, the image acquisition module includes a camera and a camera control unit;

[0086] The camera is used to capture the image of the vehicle dashboard to obtain a dashboard image;

[0087] The camera control unit is used to adjust the shooting parameters of the camera.

[0088] Specifically, the camera is responsible for capturing images of the car dashboard to ensure that the images are clear and complete. The camera control unit can automatically adjust the camera's shooting parameters (such as exposure time, focal length, ISO, etc.) according to the shooting environment (such as light conditions, dashboard distance, etc.), ensuring that high-quality dashboard images can be obtained under different conditions, improving the accuracy of subsequent indication recognition.

[0089] To further illustrate, the robot arm module includes a robot arm and a robot arm control unit;

[0090] The mechanical arm is used to move to the front of the automobile dashboard according to the position information;

[0091] The robot arm control unit is used to accurately control the movement of the robot arm.

[0092] Specifically, the robotic arm moves precisely to the front of the dashboard through mechanisms such as motor drive based on the dashboard position information provided by the image processing and analysis module. The robotic arm control unit is responsible for receiving the position information and accurately controlling the movement path and speed of the robotic arm through algorithms to ensure that the robotic arm can reach the specified position smoothly and accurately. The precise control of the robotic arm is achieved, ensuring that the image acquisition module can accurately align with the dashboard, improving the accuracy and stability of the shooting. At the same time, the automation level of the device is enhanced, the need for manual operation is reduced, and work efficiency is improved.

[0093] To further illustrate, the image processing and analysis module includes a preprocessing unit, a feature extraction unit and an indication recognition unit;

[0094] The preprocessing unit is used to preprocess the collected instrument panel image;

[0095] The feature extraction unit is used to extract key features in the instrument panel image;

[0096] The indication recognition unit is used to recognize the indication information on the vehicle dashboard according to the extracted key features.

[0097] Specifically, the preprocessing unit performs preprocessing operations such as denoising and contrast enhancement on the collected dashboard image to improve the image quality.

[0098] The feature extraction unit uses image processing algorithms (such as edge detection, contour extraction, etc.) to extract key features (such as digital area, pointer position, etc.) in the instrument panel image to reduce the difficulty of subsequent indication recognition.

[0099] The indication recognition unit uses a deep learning model to identify the indication information on the dashboard based on the extracted key features, thereby achieving accurate recognition of the dashboard indications and improving the accuracy and reliability of recognition.

[0100] A method for detecting the indication of a vehicle instrument panel, applied to the device for detecting the indication of a vehicle instrument panel as described above, comprises the following steps:

[0101] S1. Use an image acquisition module to capture an image of a car dashboard to obtain a dashboard image;

[0102] S2, according to the position information of the dashboard in the dashboard image, control the robot arm module to move to the front of the car dashboard to collect the image;

[0103] S3, preprocessing the collected dashboard image to extract key features of the pointer in the dashboard image;

[0104] S4, executing the instrument panel indication detection algorithm according to the extracted key features to identify the indication information on the automobile instrument panel;

[0105] S5. Output the recognized instrument panel indication.

[0106] The method for detecting the indication of a car dashboard of the present invention is mainly applied to a specific detection device. The workflow of the method is meticulous and efficient, and the specific steps are as follows:

[0107] First, the car dashboard is photographed using an image acquisition module (such as a high-definition camera) to obtain a clear dashboard image (S1). This step is the basis for subsequent processing, and ensuring image quality is crucial for accurately identifying dashboard readings.

[0108] Then, according to the specific position information of the dashboard in the dashboard image, the control system accurately controls the robot arm module to move to the front of the car dashboard, so as to perform more detailed and accurate image acquisition (S2). The precise positioning capability of the robot arm ensures the accuracy and stability of image acquisition.

[0109] Then, the collected dashboard image is preprocessed, and this step aims to extract the key features of the pointer in the dashboard image, such as the position, shape and angle of the pointer (S3). These key features are crucial for subsequent indication recognition.

[0110] Subsequently, based on the extracted key features, the dashboard indication detection algorithm is executed. The algorithm uses advanced image processing technology and machine learning algorithms to accurately identify the indication information on the car dashboard, such as speed, fuel level, water temperature, etc. (S4).

[0111] Finally, the identified instrument panel indications are output and may be displayed on a screen or transmitted to other systems or devices for further processing (S5).

[0112] In summary, the method of the present invention has the following beneficial effects:

[0113] 1. High efficiency: The precise positioning of the robotic arm module and the rapid shooting of the image acquisition module greatly improve the efficiency of instrument panel indication detection.

[0114] 2. Accuracy: Using advanced image processing technology and machine learning algorithms, it can accurately extract key features in dashboard images, accurately identify reading information, and reduce the misjudgment rate.

[0115] 3. Automation: The entire testing process is highly automated, reducing manual intervention and labor costs while improving the objectivity and consistency of testing.

[0116] 4. Wide applicability: It can be applied to various types and models of car dashboards, with strong versatility and flexibility.

[0117] To further illustrate, in S3, the preprocessing of the collected dashboard image specifically includes the following steps:

[0118] First, the collected dashboard image is preprocessed, including using color filtering and feature detection image processing technology to find the car dashboard area and extract the image of the area. This can help focus on the car dashboard and reduce the complexity of subsequent processing.

[0119] The purpose of color filtering is to extract the dashboard area by identifying a specific color range in the image. Dashboards usually have specific colors, such as white, black, red, green, etc., so we can filter the image by setting the color range. The HSV (hue, saturation, brightness) color space is usually selected instead of RGB because HSV is easier to handle hue, saturation, and brightness independently, which is convenient for color selection.

[0120] S31. Set the range of hue H, saturation S and brightness V in the dashboard image, and filter by the following conditions:

[0121] H min ≤H≤H max

[0122] S min ≤S≤S max

[0123] V min ≤V≤V max

[0124] Among them, H min , H max , S min , S max 、V min and V max By presetting;

[0125] Among them, the condition for extracting red is set as:

[0126] H min =0°, H max =20°

[0127] S min =0.5, S max =1.0

[0128] V min =0.3, V max =1.0

[0129] Thus, the red area in the dashboard image that meets the above conditions is filtered out;

[0130] Since there will be some errors and uncertainties in the process of color filtering, they may come from noise, lighting changes, color similarity and other factors. Therefore, it is necessary to suppress noise and use denoising methods such as Gaussian blur or median filtering to smooth the image and reduce the impact of noise on edge detection and color filtering.

[0131] S32. Use Gaussian blur formula for denoising:

[0132]

[0133] Among them, ω(i,j) represents the Gaussian weight, k represents the window size, and I(x,y) represents the original image;

[0134] Since color filtering may cause inaccurate selection of the hue range, the error can be handled by expanding the color range to make the filtering more robust, such as appropriately increasing or decreasing the threshold range of hue, saturation, and brightness.

[0135] S33, adopting a fault tolerance mechanism to appropriately increase or decrease the threshold range of hue, saturation and brightness;

[0136] Since the size of features may vary in different images, multi-scale detection is effective. Specifically, feature detection through image pyramids or multi-level images can ensure that the detected areas are valid at different scales.

[0137] S34. According to the preset scaling factor, the image pyramid formula is used to perform feature detection to generate a dashboard image of a preset scale:

[0138]

[0139] Among them, s is the scaling factor used to generate images of different scales.

[0140] To further illustrate, in S3, extracting the key features of the pointer in the dashboard image includes obtaining edge information of the car dashboard using an edge detection algorithm, which specifically includes the following steps:

[0141] S35, applying the Canny edge detection operator to the dashboard image of the preset size captured in step S34 to accurately obtain edge information in the image, which helps to more accurately identify the shape and structure of the automobile dashboard.

[0142] Since the car dashboard usually has geometric features (such as circle, rectangle, pointer, etc.), these geometric areas can be further extracted through edge detection algorithms.

[0143] S35. Apply the Canny edge detection algorithm to the dashboard image of the preset scale to obtain edge information:

[0144] Set the gradient amplitude of the image G(x,y), let G x and G y is the gradient of the image in the horizontal and vertical directions, and the relationship between the two is:

[0145]

[0146] S36. According to the gradient amplitude G(x,y) of the image, a high threshold high_threshol and a low threshold low_threshold are applied to distinguish edge information:

[0147] When G(x,y)>high_threshol, it is considered a strong edge;

[0148] When low_threshold≤G(x,y)≤high_threshold, it is considered a weak edge;

[0149] When G(x,y)<low_threshold, it is considered as non-edge;

[0150] It should be noted that the high threshold and the low threshold need to be set reasonably, otherwise: when the high threshold is too high, only very strong edges will be detected, and weaker edges may be missed. When the low threshold is too low, more false detections and false edges may occur.

[0151] Adjust the ratio between the low threshold and the high threshold through experimentation. Usually, it is a common practice to set the low threshold to 1 / 3 to 1 / 2 of the high threshold when initially defining it. For example, when the high threshold is set to 100, the low threshold can be set to 30 or 50.

[0152] To further illustrate, in S3, extracting the key features of the pointer in the dashboard image includes obtaining the periphery of the car dashboard using a circle detection algorithm, which specifically includes the following steps:

[0153] Hough transform is used to find the circular structure on the dashboard, which represents the periphery of the dashboard. In this way, the overall shape and position of the dashboard can be determined, providing a basis for subsequent pointer detection.

[0154] S37, using the edge information of the image extracted in step S36 as input, using Hough transform to find the circular structure on the dashboard, representing the periphery of the dashboard:

[0155] Define the peripheral equation of the instrument panel as: (xa) 2 +(yb) 2 =r2

[0156] Among them, (a, b) represents the coordinates of the center of the circle, and r represents the radius of the circle;

[0157] Through the Hough transform, we can accumulate votes in the parameter space and finally detect the circles that meet the conditions. The application steps of the circular Hough transform are:

[0158] S38. For each edge point (x, y) in the image, consider all circles with different radii r to which it belongs; enumerate different circle center coordinates a, b and radii r, and record the voting results of each group (a, b, r);

[0159] S39, each time the edge points in the image are traversed, the circle accumulator is updated according to the circle parameter space, and the votes of different circle centers and radii are recorded; if multiple edge points vote for the same circle parameter combination, it indicates that the circle exists in the image;

[0160] S310, finding the edge point with the most votes in the circular accumulator, which represents the most likely circular structure in the image, that is, the periphery of the automobile dashboard.

[0161] In order to filter out accurate circles, reasonable parameters need to be set:

[0162] Minimum and maximum radius: Set a reasonable radius range to avoid detecting irrelevant circles. For example, if detecting a dial on a dashboard, you might set the radius between 50 and 200 pixels.

[0163] Threshold: Select the circle with the most votes by setting a threshold. For example, a circle is considered valid only when the number of votes for the circle exceeds the preset threshold.

[0164] Accumulator resolution: The resolution of the parameter space determines the accuracy of the Hough transform. If the resolution is too low, the circle may not be detected accurately; if the resolution is too high, the amount of calculation increases and it is easy to be interfered by noise. In general, a resolution of 1-2 pixels is selected for search.

[0165] Screening test results:

[0166] Filtering of circle size: Filter out possible circles by setting the radius range. For dashboard images, the radius of the background circle is usually larger, while the radius of the dial circle is smaller, so different circular structures can be distinguished based on the size of the radius.

[0167] Voting number screening: By setting a voting threshold, only circles with voting numbers higher than the threshold are retained. This can effectively filter out false circles caused by noise or other reasons.

[0168] Position filtering: By checking the coordinates of the center of the circle, you can further filter out the main circular structures of the dashboard, such as the background circle, dial circle, etc.

[0169] To further illustrate, in S3, extracting the key features of the pointer in the dashboard image includes obtaining the pointer of the car dashboard using a line detection algorithm, which specifically includes the following steps:

[0170] Hough transform is further used to find the lines on the dashboard, especially the lines of the pointer. By detecting these lines, the position and direction of the pointer can be accurately located.

[0171] S311, using the edge information of the image extracted in step S36 as input, using Hough transform to find the straight line structure on the dashboard, representing the pointer of the dashboard, and using two-dimensional parameters ρ and θ in the parameter space to represent the equation of the straight line:

[0172] ρ=xcos(θ)+ysin(θ)

[0173] Where ρ is the shortest distance from the origin to the line, and θ is the angle between the line and the x-axis;

[0174] Since the basic process of the line detection algorithm is similar to circle detection, the line with the most votes is found in the parameter space through the voting of edge points:

[0175] S312, for each edge point (x, y) in the image, calculate the parameters ρ and θ of all the straight lines to which it belongs;

[0176] S313, voting the (ρ, θ) of each edge point to the corresponding position in the parameter space, forming a linear accumulator, and recording different (ρ, θ) combinations;

[0177] S314, finding the (ρ, θ) combination with the most votes in the line accumulator, which represents the most likely straight line structure in the image, that is, the pointer of the car dashboard.

[0178] In order to filter out accurate lines, reasonable parameters need to be set:

[0179] Resolution: Set the appropriate parameter spatial resolution to ensure the accuracy of line detection. Generally speaking, higher resolution can provide more details, but also increase the computational complexity.

[0180] Threshold: Set a voting threshold to select detected lines. For example, only lines with votes exceeding a certain threshold are considered valid.

[0181] Screening test results:

[0182] Line length: You can filter out important lines based on their length or vote count. For dashboards, needles are usually long straight lines, while tick marks are shorter.

[0183] Direction screening: By checking the direction of the line (i.e. the value of θ), you can further screen out the straight lines that are consistent with the pointer direction. The direction of the instrument panel pointer is generally close to a certain angle (such as a horizontal line or a vertical line), and the threshold can be used to select the straight lines that meet the conditions.

[0184] To further illustrate, in S4, executing the instrument panel indication detection algorithm based on the extracted key features to identify the indication information on the vehicle instrument panel specifically includes the following steps:

[0185] Specifically, the reading of the instrument panel is calculated according to the angle between the pointer and the vertical direction; this step can determine the current reading value according to the position and direction of the pointer and the scale information of the instrument panel.

[0186] S41, the equation of the straight line obtained by line detection: ρ = xcos (θ) + ysin (θ) to obtain the angle θ′ between the pointer and the vertical direction;

[0187] S42, mapping the relationship between the angle θ′ and the scale of the car dashboard, calculating the proportional relationship between the angle and the reading, and obtaining a corresponding mathematical formula.

[0188] The above process requires determining the reading range of the instrument panel, understanding the number of scales on the instrument panel and how they are distributed; it is necessary to pay attention to whether the angle range of the instrument panel is 0° to 360° and whether the starting scale of the instrument panel is 0. It is also necessary to pay attention to whether the scales of the instrument panel are evenly distributed. If they are not evenly distributed (for example, a pressure gauge or a speedometer), nonlinear mapping needs to be performed according to the specific instrument design. This may require calibration in advance by sampling or manually entering scale information.

[0189] In practical applications, some special situations may occur in dashboard images, such as pointer occlusion, blurred scale, etc., which require additional processing:

[0190] Handling pointer occlusion: Pointer occlusion refers to the situation where the pointer is partially or completely blocked by other objects (such as scales, the pointer itself, or external objects), making it impossible to clearly extract the pointer position. Methods for handling occlusion problems include:

[0191] Image preprocessing: Use image smoothing and denoising techniques to enhance the contrast of the pointer in the image and reduce the impact of the occluded part.

[0192] Multiple image sampling: Take multiple shots of the same dashboard at different angles or lighting conditions, and use multiple images to average the pointer position to reduce errors caused by occlusion.

[0193] Inference based on prior knowledge: By assuming that the pointer is usually located at a certain position, the angle of the pointer is inferred by combining other information (such as scale, background circle, etc.).

[0194] Dealing with scale blur: Scale blur usually occurs when the scale numbers or lines on the instrument panel become difficult to recognize due to poor image quality or lighting problems. Solutions include:

[0195] Enhance image clarity: Use image sharpening technology to improve the clarity of scale lines.

[0196] Template matching: If the scale is fixed, template matching techniques can be used to find and match specific scale patterns in the image.

[0197] Lighting balance: Use image lighting balance technology to remove blurring caused by lighting changes and make the image more uniform.

[0198] The technical principle of the present invention is described above in conjunction with specific embodiments. These descriptions are only for explaining the principle of the present invention and cannot be interpreted as limiting the scope of protection of the present invention in any way. Based on the explanations herein, those skilled in the art can associate other specific embodiments of the present invention without creative work, and these equivalent variations or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A device for detecting the indication of a car instrument panel, characterized in that: include: An image acquisition module is used to capture an image of a car dashboard to obtain an image of the dashboard; A mechanical arm module, connected to the image acquisition module, and used to move to the front of the vehicle dashboard to acquire an image according to the position information of the vehicle dashboard in the dashboard image taken by the image acquisition module; An image processing and analysis module, used for receiving the dashboard image acquired by the image acquisition module, and executing a dashboard indication detection algorithm to identify indication information on the vehicle dashboard; The output module is used to output the identified vehicle dashboard indication.

2. A device for detecting the indication of a car instrument panel according to claim 1, characterized in that: The image acquisition module includes a camera and a camera control unit; The camera is used to capture the image of the vehicle dashboard to obtain a dashboard image; The camera control unit is used to adjust the shooting parameters of the camera.

3. The device for detecting the indication of the automobile instrument panel according to claim 1, characterized in that: The robotic arm module includes a robotic arm and a robotic arm control unit; The mechanical arm is used to move to the front of the automobile dashboard according to the position information; The robotic arm control unit is used to accurately control the movement of the robotic arm.

4. The device for detecting the indication of the automobile instrument panel according to claim 1, characterized in that: The image processing and analysis module includes a preprocessing unit, a feature extraction unit and an indication recognition unit; The preprocessing unit is used to preprocess the collected instrument panel image; The feature extraction unit is used to extract key features in the instrument panel image; The indication recognition unit is used to recognize the indication information on the vehicle dashboard according to the extracted key features.

5. A method for detecting the indication of a car dashboard, characterized in that: A device for detecting the indication of a vehicle instrument panel as claimed in any one of claims 1 to 4, comprising the following steps: S1. Use an image acquisition module to capture an image of a car dashboard to obtain a dashboard image; S2, according to the position information of the dashboard in the dashboard image, control the robot arm module to move to the front of the car dashboard to collect the image; S3, preprocessing the collected dashboard image to extract key features of the pointer in the dashboard image; S4, executing the instrument panel indication detection algorithm according to the extracted key features to identify the indication information on the automobile instrument panel; S5. Output the recognized instrument panel indication.

6. A method for detecting vehicle instrument panel indications according to claim 1, characterized in that: In S3, the collected dashboard image is preprocessed, specifically including the following steps: S31. Set the range of hue H, saturation S and brightness V in the dashboard image, and filter by the following conditions: H min ≤H≤H max S min ≤S≤S max In min ≤V≤V max Among them, H min , H max , S min , S max 、V min and V max By presetting; S32. Use Gaussian blur formula for denoising: Among them, ω(i,j) represents the Gaussian weight, k represents the window size, and I(x,y) represents the original image; S33, adopting a fault tolerance mechanism to appropriately increase or decrease the threshold range of hue, saturation and brightness; S34. According to the preset scaling factor, the image pyramid formula is used to perform feature detection to generate a dashboard image of a preset scale: Among them, s is the scaling factor used to generate images of different scales.

7. A method for detecting vehicle instrument panel indications according to claim 6, characterized in that: In S3, extracting the key features of the pointer in the dashboard image includes acquiring edge information of the car dashboard using an edge detection algorithm, which specifically includes the following steps: S35. Apply the Canny edge detection algorithm to the dashboard image of the preset scale to obtain edge information: Set the gradient amplitude of the image G(x,y), let G x and G y is the gradient of the image in the horizontal and vertical directions, and the relationship between the two is: S36. According to the gradient amplitude G(x,y) of the image, a high threshold high_threshol and a low threshold low_threshold are applied to distinguish edge information: When G(x,y)>high_threshol, it is considered a strong edge; When low_threshold≤G(x,y)≤high_threshold, it is considered a weak edge; When G(x,y)<low_threshold, it is considered as a non-edge.

8. A method for detecting vehicle instrument panel indications according to claim 7, characterized in that: In S3, extracting the key features of the pointer in the dashboard image includes obtaining the periphery of the car dashboard using a circle detection algorithm, which specifically includes the following steps: S37, using the edge information of the image extracted in step S36 as input, using Hough transform to find the circular structure on the dashboard, representing the periphery of the dashboard: Define the peripheral equation of the instrument panel as: (xa) 2 +(yb) 2 =r 2 Among them, (a, b) represents the coordinates of the center of the circle, and r represents the radius of the circle; S38. For each edge point (x, y) in the image, consider all circles with different radii r to which it belongs; enumerate different circle center coordinates a, b and radii r, and record the voting results of each group (a, b, r); S39, each time the edge points in the image are traversed, the circle accumulator is updated according to the circle parameter space, and the votes of different circle centers and radii are recorded; if multiple edge points vote for the same circle parameter combination, it indicates that the circle exists in the image; S310, finding the edge point with the most votes in the circular accumulator, which represents the most likely circular structure in the image, that is, the periphery of the automobile dashboard.

9. A method for detecting vehicle instrument panel indications according to claim 8, characterized in that: In S3, extracting the key features of the pointer in the dashboard image includes obtaining the pointer of the car dashboard using a line detection algorithm, which specifically includes the following steps: S311, using the edge information of the image extracted in step S36 as input, using Hough transform to find the straight line structure on the dashboard, representing the pointer of the dashboard, and using two-dimensional parameters ρ and θ in the parameter space to represent the equation of the straight line: ρ=xcos(θ)+ysin(θ) Where ρ is the shortest distance from the origin to the line, and θ is the angle between the line and the x-axis; S312, for each edge point (x, y) in the image, calculate the parameters ρ and θ of all the straight lines to which it belongs; S313, voting the (ρ, θ) of each edge point to the corresponding position in the parameter space, forming a linear accumulator, and recording different (ρ, θ) combinations; S314, finding the (ρ, θ) combination with the most votes in the line accumulator, which represents the most likely straight line structure in the image, that is, the pointer of the car dashboard.

10. A method for detecting vehicle instrument panel indications according to claim 9, characterized in that: In S4, the method of executing the instrument panel indication detection algorithm based on the extracted key features to identify the indication information on the vehicle instrument panel specifically includes the following steps: S41, the equation of the straight line obtained by line detection: ρ = xcos (θ) + ysin (θ) to obtain the angle θ′ between the pointer and the vertical direction; S42, mapping the relationship between the angle θ′ and the scale of the car dashboard, calculating the proportional relationship between the angle and the reading, and obtaining a corresponding mathematical formula.

Images