Intelligent golf ball motion camera with pitching and horizontal automatic rotation functions
By designing a golf intelligent sports camera with pitch and horizontal automatic rotation, the problem of human control in the existing technology is solved, and the camera automatically tracks players and balls, improving the user experience.
Patent Information
- Application Number
- CN202411627513.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-14
- Publication Date
- 2025-02-28
AI Technical Summary
The existing golf sports recording tools require human control and cannot automatically track players and balls, resulting in poor user experience.
A golf ball intelligent sports camera with pitch and horizontal automatic rotation is designed, using an adjustable gimbal and battery. The camera body can automatically rotate horizontally and pitch, and has the function of automatically identifying players and balls.
The camera automatically tracks players and balls, improving the user experience, and users can easily take photos and videos on the golf course.
Smart Images
Figure CN119532581A_ABST
Abstract
Description
[Technical field]
[0001] The invention relates to the technical field of sports cameras, and in particular to an intelligent golf sports camera with automatic pitch and horizontal rotation. [Background technology]
[0002] In the golf industry, most golfers currently use mobile phones or cameras to record their sports videos or scenes at the golf course. These tools used in the existing technology require manual control to achieve, or the recording effect is relatively poor. These tools cannot be automatically operated, are inconvenient to use, and cannot track the trajectory of the ball, which affects the user experience. [Summary of the invention]
[0003] In order to overcome the above problems, the present invention proposes a golf smart sports camera with automatic pitch and horizontal rotation which can effectively solve the above problems.
[0004] The present invention provides a technical solution to solve the above technical problems: a golf smart sports camera with automatic pitch and horizontal rotation is provided, comprising a camera body, a gimbal and a fixed rod, wherein the camera body is connected to the gimbal, and the gimbal is connected to the fixed rod; the gimbal comprises a first connecting arm and a second connecting arm, wherein the first connecting arm and the second connecting arm are connected via a damping assembly, one end of the first connecting arm is connected to the camera body, and one end of the second connecting arm is connected to the fixed rod; a pitch rotation motor is provided at the connection between the first connecting arm and the camera body, and a horizontal rotation motor is provided at the connection between the second connecting arm and the fixed rod; the horizontal rotation motor can drive the gimbal to rotate 360° horizontally, and the pitch rotation motor can drive the camera body to rotate 360° in pitch.
[0005] Preferably, the camera body has the functions of taking pictures, recording videos, and tracking objects or golf balls or clubs.
[0006] Preferably, the fixed rod is fixed in the venue. After powering on, the camera body is calibrated with the fixed rod to automatically identify the athlete or player. Then, the ball is used as the target and the camera body tracks the ball. When the ball moves in a parabolic motion, the camera body also tracks the player and the ball to record and take pictures.
[0007] Preferably, the camera body includes a first camera and a second camera, the first camera and the second camera are both arranged at the front end of the camera body, the first camera is located above the second camera, and the first camera is a wide-angle camera.
[0008] Preferably, the fixing rod comprises a control part, a control circuit board is arranged in the control part, and the pitch rotation motor and the horizontal rotation motor are both connected to the control circuit board.
[0009] Preferably, a camera battery is arranged in the camera body, and a backup battery is arranged in the fixing rod. When the battery in the camera body is used up, the backup battery in the fixing rod can provide power for the camera body.
[0010] Preferably, the damping assembly comprises a damping shaft, and the relative position between the first connecting arm and the second connecting arm can be manually adjusted.
[0011] Preferably, a power button is disposed on the outer side of the control unit, and the power button is connected to a control circuit board.
[0012] Preferably, a WIFI button and a USB interface are provided on the outside of the control unit, and both the WIFI button and the USB interface are connected to the control circuit board.
[0013] Preferably, the camera body comprises a detection module inside, and a highly robust golf ball detection algorithm is integrated in the detection module.
[0014] Compared with the prior art, the golf smart sports camera with automatic pitch and horizontal rotation of the present invention is equipped with an adjustable pan head and a battery, and can automatically rotate horizontally and in pitch. The camera body 1 can automatically capture, automatically track the ball, and track the player, and has the function of automatically identifying the ball and the player. A person can take photos and record videos while playing golf. It is easy to use, solves the pain points of users, and is conducive to improving the user experience.
Brief Description of the Drawings
[0015] Figure 1 The first stereoscopic image of the golf intelligent sports camera with automatic pitch and horizontal rotation of the present invention;
[0016] Figure 2 A second stereoscopic image of the golf intelligent sports camera with automatic pitch and horizontal rotation according to the present invention;
[0017] Figure 3 This is a third stereoscopic view of the golf smart sports camera with automatic pitch and horizontal rotation according to the present invention. [Specific implementation method]
[0018] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and implementation examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0019] It should be noted that in the embodiments of the present invention, all directional indications (such as up, down, left, right, front, back, etc.) are limited to relative positions on the specified view, rather than absolute positions.
[0020] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0021] See also Figures 1 to 3 The golf smart sports camera with automatic pitch and horizontal rotation of the present invention comprises a camera body 1, a pan head 2 and a fixing rod 3, wherein the camera body 1 is connected to the pan head 2, and the pan head 2 is connected to the fixing rod 3.
[0022] The gimbal 2 includes a first connecting arm 21 and a second connecting arm 22 , which are connected via a damping assembly 23 , one end of the first connecting arm 21 is connected to the camera body 1 , and one end of the second connecting arm 22 is connected to the fixing rod 3 .
[0023] A pitch motor 6 is provided at the connection between the first connecting arm 21 and the camera body 1, and a horizontal motor 7 is provided at the connection between the second connecting arm 22 and the fixing rod 3. The horizontal motor 7 can drive the gimbal 2 to rotate 360° horizontally, and the pitch motor 6 can drive the camera body 1 to rotate 360° in pitch.
[0024] The camera body 1 has the functions of taking photos, recording videos, tracking objects or golf balls and clubs. The fixed rod 3 is fixed in the field. After the camera body 1 is turned on, the fixed rod 3 is used for calibration to automatically identify the athlete or golfer. Then the ball is used as the target to let the camera body 1 track the ball. When the ball moves in a parabolic motion, the camera body 1 also tracks the golfer and the ball to record and take photos. The camera body 1 can rotate 360° horizontally and 360° in pitch through the pan / tilt 2, so that the camera body 1 can shoot and record the players, golfers and balls without blind spots.
[0025] The camera body 1 includes a first camera 11 and a second camera 12 . The first camera 11 and the second camera 12 are both arranged at the front end of the camera body 1 . The first camera 11 is located above the second camera 12 . The first camera 11 is a wide-angle camera.
[0026] The fixing rod 3 includes a control unit 4 , in which a control circuit board is disposed. The pitch motor 6 and the horizontal rotation motor 7 are both connected to the control circuit board.
[0027] The camera body 1 is provided with a camera battery, and the fixing rod 3 is provided with a 10000W backup battery. When the battery in the camera body 1 is used up, the backup battery in the fixing rod 3 can provide power for the camera body 1, thus meeting the players' unexpected needs.
[0028] The damping assembly 23 includes a damping shaft, and the relative position between the first connecting arm 21 and the second connecting arm 22 can be manually adjusted.
[0029] A power button 5 is disposed outside the control unit 4 , and the power button 5 is connected to a control circuit board.
[0030] A WIFI button 8 and a USB interface 9 are disposed on the outer side of the control unit 4 , and both the WIFI button 8 and the USB interface 9 are connected to the control circuit board.
[0031] The golf smart sports camera with automatic pitch and horizontal rotation of the present invention is equipped with an adjustable pan head 2 and a battery, and can automatically rotate horizontally and in pitch. The camera body 1 can automatically capture, automatically track the ball, and track the player, and has the function of automatically identifying the ball and the player. A person can also take photos and record videos while playing golf. The camera is easy to use, solves the pain points of users, and is conducive to improving the user experience.
[0032] The golf ball intelligent sports camera with automatic pitch and horizontal rotation of the present invention includes a detection module inside the camera body 1, and a highly robust golf ball detection algorithm is integrated in the detection module. The highly robust golf ball detection algorithm can perform stable detection in strong light and weak light environments and adapt to various lighting conditions; it has high detection accuracy, and can detect golf balls at a long distance by fusing the bottom layer and the upper layer through multi-scale feature extraction, thereby improving the accuracy of golf ball detection; it has good occlusion processing capabilities, and can still accurately detect golf balls even if they are partially occluded.
[0033] Golf ball detection is a key technology in the monitoring and training systems of golf courses. Existing golf ball detection methods do not work well in complex environments such as strong light, weak light, long distance and occlusion, resulting in low detection accuracy. The highly robust golf ball detection algorithm integrated in the detection module is highly robust to changes in illumination, has high accuracy for long-distance golf ball detection, and can detect partially occluded golf balls.
[0034] The highly robust golf ball detection algorithm comprises the following steps:
[0035] Step S1, model training.
[0036] The step S1 includes the following steps:
[0037] Step S11, data collection: collect golf ball images under different lighting conditions, different distances and partial occlusions to build a diverse data set. The so-called different lighting conditions are divided into strong light and weak light; the so-called different distances are divided into near, medium and far; the so-called partial occlusion mainly refers to the golf club head partially occluding the golf ball.
[0038] Data collection. The data collection phase collects diverse data. During the specific implementation process, the preparation of the data set includes the following aspects:
[0039] (1) Lighting conditions:
[0040] Strong light environment: Golf ball images shot under strong direct sunlight, simulating the environment of a golf course at noon.
[0041] Low-light environment: Golf ball images taken at sunrise, sunset or cloudy conditions, simulating golf course environments in low-light conditions.
[0042] Night environment: Golf ball images shot at night or under lighting, simulating nighttime competition or training scenarios.
[0043] (2) Distance change:
[0044] Close Distance: Images taken with the golf ball closer to the camera, ensuring that the golf ball occupies a larger area of the image.
[0045] Medium Distance: Images taken with the golf ball at a moderate distance from the camera, simulating the viewing distance during regular play or practice.
[0046] Long distance: The image is taken when the golf ball is far away from the camera. The golf ball is very small in the image, and the algorithm's performance in long-distance detection is tested.
[0047] (3) Occlusion situation:
[0048] Partial Occlusion: An image of a golf ball partially obscured by a club, grass, or other objects, testing the robustness of the algorithm in the presence of occlusion.
[0049] Fully Exposed: An image where the golf ball is fully exposed in the field of view, ensuring the algorithm can accurately detect it without obstructions.
[0050] Complex background: Golf ball images taken against complex backgrounds (such as woods, spectator seats) to test the algorithm’s detection capabilities in the presence of background interference.
[0051] (4) Environmental factors:
[0052] Weather changes: Golf ball images taken under different weather conditions, such as sunny, cloudy, rainy, and foggy days, to ensure the algorithm’s adaptability in all weather conditions.
[0053] Turf State: Golf ball images taken in different turf states (e.g., neatly trimmed lawn, overgrown grass, or sandy ground) to test the algorithm’s performance under different ground conditions.
[0054] Step S12, image preprocessing: perform illumination enhancement, image denoising, and data enhancement on the image in step S11. The so-called illumination enhancement is to enhance the image contrast and improve the image visibility; the so-called image denoising is to use a filtering algorithm to denoise the image. The filtering algorithm mainly sets a suitable convolution kernel and performs convolution operation with the image to eliminate image noise and improve image quality; the so-called data enhancement is mainly rotation, flipping, zooming, and translation operations. The image is randomly rotated at a certain angle to simulate golf balls at different angles; the image is randomly flipped horizontally or vertically; the diversity of the data set is increased; the image is randomly scaled to simulate golf balls at different distances; the image is randomly translated to simulate the change in viewing angle caused by the change in camera position.
[0055] Image preprocessing. Image enhancement, image denoising and data enhancement are performed in the data acquisition stage. In the specific implementation process, image enhancement uses adaptive histogram equalization (CLAHE). Adaptive histogram equalization (CLAHE) is a technology for enhancing image contrast, which is particularly suitable for image processing under different lighting conditions. Specific steps: convert the image into a grayscale image; divide the grayscale image into several small areas (tiles); perform histogram equalization on each small area to enhance the local contrast; use bilinear interpolation to smooth the boundaries between small areas and synthesize the enhanced image; merge the enhanced grayscale image with the original color image to generate an enhanced color image. After these steps, the visibility of the image in strong and weak light environments can be effectively improved, making the golf ball easier to detect under complex lighting conditions. In the specific implementation process, image denoising uses the Gaussian filtering algorithm. Opencv is a powerful image operation library. Opencv provides the GaussianBlur interface to complete the image Gaussian filtering. After setting the standard deviation parameters, call GaussianBlur to complete the denoising of the input image. During the collective implementation, data enhancement was mainly completed using the Pytorch framework, which is an open source deep learning framework that provides dynamic computational graphs, flexible neural network building blocks, and powerful GPU acceleration capabilities, allowing researchers and developers to easily build, train, and deploy deep learning models. Pytorch provides a transforms object, which is used to perform rotation, flipping, zooming, and translation operations.
[0056] Step S13, model training: Use the images from the image preprocessing stage to train a network model with multi-scale feature extraction capabilities. The so-called multi-scale feature extraction is to extract data features from different scales (or resolutions), which can not only extract high-level semantic information, but also low-level semantic information. Fusion of features at different scales can improve the robustness and stability of the features, so that the algorithm can handle target objects of different sizes and resolutions.
[0057] Model training. Train a network model with multi-scale feature extraction capabilities. In the specific implementation process, select the yolov5 model structure. The yolov5 model contains the PANet structure, which includes FPN (Feature Pyramid Network) and PAN (Path Aggregation Network). FPN performs top-down feature fusion through multi-scale feature maps extracted from Backbone, and fuses high-resolution low-level feature maps with low-resolution high-level feature maps to capture more details and contextual information. PAN adds bottom-up path aggregation on the basis of FPN. Bottom-up path aggregation enhances the representation ability of feature maps by transferring the bottom feature map to the top, which helps to improve the detection effect of small targets.
[0058] Step S2, edge reasoning.
[0059] The step S2 includes the following steps:
[0060] Step S21, image input: a high dynamic range image is captured from an HDR camera. HDR is an image processing technology used to capture and display a wider range of brightness in a scene. Traditional image and video technology (SDR, Standard Dynamic Range) can only capture and display a limited range of brightness, while HDR technology can better preserve the details of the brightest and darkest areas of the scene, making the image closer to the real world seen by the human eye.
[0061] Image input. Use the Sony IMX678 image sensor to obtain high dynamic range images. The Sony IMX678 image sensor can obtain clear images regardless of strong light, weak light, front light, back light, or during the day.
[0062] Step S22, image preprocessing: image enhancement and image denoising. Image enhancement plays a vital role in the field of computer vision and image processing. By improving the quality and visibility of the image, subsequent analysis and processing are more accurate and effective. Some commonly used image enhancement techniques include histogram equalization, gamma correction sharpening, etc. Image denoising is an important task in image processing, which aims to remove noise from the image while retaining the details and features of the image as much as possible. Common image denoising methods include mean filtering, median filtering, Gaussian filtering, autoregressive moving average model denoising, etc.
[0063] Image preprocessing. Perform image enhancement, image denoising, and data enhancement on the high dynamic range images acquired by the Sony IMX678 image sensor. In the specific implementation process, image enhancement uses adaptive histogram equalization (CLAHE), which is a technology for enhancing image contrast and is particularly suitable for image processing under different lighting conditions. Specific steps: convert the image into a grayscale image; divide the grayscale image into blocks and divide the image into several small areas (tiles); perform histogram equalization on each small area to enhance local contrast; use bilinear interpolation to smooth the boundaries between small areas and synthesize the enhanced image; merge the enhanced grayscale image with the original color image to generate an enhanced color image. After these steps, the visibility of the image in strong and weak light environments can be effectively improved, making the golf ball easier to detect under complex lighting conditions. In the specific implementation process, image denoising uses the Gaussian filtering algorithm. Opencv is a powerful image operation library. Opencv provides the GaussianBlur interface to complete image Gaussian filtering. After setting the standard deviation parameters, call GaussianBlur to complete the denoising of the input image.
[0064] Step S23, golf ball detection: Use a neural network model to detect the golf ball and output a rectangular box of the golf ball position. The so-called neural network is a computational model that simulates the biological nervous system and is used to process complex data and perform pattern recognition. They are composed of a large number of neurons (or nodes), which are interconnected by weights to form a network structure. Neural networks can automatically learn complex mapping relationships from input to output and are widely used in image recognition, natural language processing, speech recognition and other fields. Common types of neural networks include convolutional neural networks (CNN), recurrent neural networks (RNN), Transformer, etc. Convolutional neural networks extract spatial features of images through convolutional layers, pooling layers and fully connected layers. Convolutional layers are used for local perception, and pooling layers are used for dimensionality reduction and reduction of computational complexity. They are used for image processing and computer vision, such as image classification, target detection, image segmentation, etc. Recurrent neural networks (RNN) have cyclic connections and can process sequence data. RNN saves information of previous time steps through hidden states and is mainly used to process sequence data, such as time series prediction, speech recognition, natural language processing, etc. Transformer is a neural network based on the self-attention mechanism, which can process sequence data in parallel. It is mainly composed of an encoder and a decoder, and extracts features through a self-attention layer and a feedforward network. It is mainly used for natural language processing, such as machine translation, text generation, and question-answering systems.
[0065] Golf ball detection. Convert the trained yolov5 model to DLC format through the snpe-pytorch-to-dlc tool provided by the SNPE SDK, quantize the model to INT8 through snpe-dlc-quant, push the quantized DLC model and SNPE dynamic library to the device through adb, write the yolov5 pre-processing, post-processing and inference code, encapsulate these codes into a gstreamer plug-in, and use the encapsulated plug-in to obtain the camera data and call the yolov5 model to detect golf balls.
[0066] Compared with the prior art, the golf smart sports camera with automatic pitch and horizontal rotation of the present invention is equipped with an adjustable pan head 2 and a battery, and can automatically rotate horizontally and in pitch. The camera body 1 can automatically capture, automatically track the ball, and track the player, and has the function of automatically identifying the ball and the player. A person can take photos and record videos while playing golf. It is easy to use, solves the pain points of users, and is conducive to improving the user experience.
[0067] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any modifications, equivalent substitutions and improvements made within the concept of the present invention should be included in the patent protection scope of the present invention.
Claims
1. A golf smart sports camera with automatic pitch and horizontal rotation, characterized in that: It includes a camera body, a pan head and a fixing rod, wherein the camera body is connected to the pan head, and the pan head is connected to the fixing rod; The gimbal comprises a first connecting arm and a second connecting arm, wherein the first connecting arm and the second connecting arm are connected via a damping assembly, one end of the first connecting arm is connected to the camera body, and one end of the second connecting arm is connected to the fixing rod; A pitch rotation motor is provided at the connection between the first connecting arm and the camera body, and a horizontal rotation motor is provided at the connection between the second connecting arm and the fixed rod; the horizontal rotation motor can drive the gimbal to rotate 360° horizontally, and the pitch rotation motor can drive the camera body to rotate 360° in pitch.
2. The golf smart sports camera with automatic pitch and horizontal rotation as claimed in claim 1, characterized in that: The camera body has the functions of taking photos, recording videos, and tracking objects or golf balls and clubs.
3. The golf ball intelligent sports camera with automatic pitch and horizontal rotation as claimed in claim 1, characterized in that: The fixing rod is fixed in the venue. After being turned on, the camera body is calibrated with the fixing rod to automatically identify the athlete or player. Then, the ball is used as the target and the camera body tracks the ball. When the ball moves in a parabolic motion, the camera body also tracks the player and the ball to record and take pictures.
4. The golf ball intelligent sports camera with automatic pitch and horizontal rotation as claimed in claim 1, characterized in that: The camera body includes a first camera and a second camera. The first camera and the second camera are both arranged at the front end of the camera body. The first camera is located above the second camera. The first camera is a wide-angle camera.
5. The golf ball intelligent sports camera with automatic pitch and horizontal rotation as claimed in claim 1, characterized in that: The fixing rod comprises a control part, a control circuit board is arranged in the control part, and the pitch rotation motor and the horizontal rotation motor are both connected to the control circuit board.
6. The golf ball intelligent sports camera with automatic pitch and horizontal rotation as claimed in claim 1, characterized in that: The camera body is provided with a camera battery, and the fixing rod is provided with a backup battery. When the battery in the camera body is used up, the backup battery in the fixing rod can provide power for the camera body.
7. The golf ball intelligent sports camera with automatic pitch and horizontal rotation as claimed in claim 1, characterized in that: The damping assembly comprises a damping shaft, and the relative position between the first connecting arm and the second connecting arm can be manually adjusted.
8. The golf ball intelligent sports camera with automatic pitch and horizontal rotation as claimed in claim 5, characterized in that: A power button is arranged on the outer side of the control part, and the power button is connected to the control circuit board.
9. The golf ball intelligent sports camera with automatic pitch and horizontal rotation as claimed in claim 5, characterized in that: A WIFI button and a USB interface are arranged on the outside of the control unit, and both the WIFI button and the USB interface are connected to the control circuit board.
10. The golf ball intelligent sports camera with automatic pitch and horizontal rotation as claimed in claim 1, characterized in that: The camera body includes a detection module inside, and a highly robust golf ball detection algorithm is integrated in the detection module.