A method and system for improving parking space recognition rate and parking ability in dark environments

Abstract

The present invention discloses a method and system for improving the recognition rate of parking spaces and parking capabilities in dark environments. First, the light sensor and surround-view camera synchronously collect the brightness and image of the vehicle's surroundings in real time, and then determine whether to activate the fill light. At the same time, the image collected by the surround-view camera is analyzed to form a bird's-eye view. Then, the coordinate information of the target parking space is obtained based on the bird's-eye view, and the brightness of the area where the target parking space is located is detected to determine whether the target parking space is in a dark area. If so, the fill light is used to compensate the light intensity of the target parking space and adjust the illumination angle of the fill light to ensure that the target parking space is always in the most suitable brightness environment. Finally, the coordinate information of the target parking space is integrated with the parking space information collected by the ultrasonic radar to complete the parking action. The present invention can effectively improve the recognition accuracy and recognition success rate of the automatic parking system for vehicles in dark environments, bringing customers a better parking experience.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic parking technology for automobiles, and in particular to a method and system for improving parking space recognition rate and parking capability in a dark environment. Background Art

[0002] As one of the autonomous driving technologies for low-speed vehicles, automatic parking is becoming increasingly popular and widely used. Under current technology, automatic parking systems can meet parking needs in most scenarios. The visual perception component of the automatic parking system consists of four surround-view cameras, which have high requirements for ambient light. Below a certain illumination level, the cameras cannot perceive parking space lines, which in turn affects the success rate of parking space recognition and results in a large number of visual parking spaces being unavailable. Therefore, for indoor scenarios such as underground parking lots, existing automatic systems have poor visual parking space recognition accuracy, large recognition errors, or even ineffective recognition in low-light environments. This greatly limits the use cases of automatic parking and affects the user experience. Summary of the Invention

[0003] The present invention provides a method and system for improving the parking space recognition rate and parking capability in low-light environments, which can effectively improve the recognition accuracy and recognition success rate of the automatic parking system for vehicles in low-light environments, lay a good foundation for planning parking paths during subsequent parking processes, and bring customers a better parking experience.

[0004] The technical solutions of the present invention are as follows:

[0005] A method for improving parking space recognition rate and parking capability in low-light environments, comprising the following steps:

[0006] S1. After the vehicle's automatic parking function is turned on, the vehicle's ambient light brightness is monitored in real time by the light sensor, and the surround-view camera collects images of the vehicle's surroundings in real time;

[0007] S2. The automatic parking controller determines whether to turn on the fill light based on the collected ambient light brightness, and processes the collected image to obtain a bird's-eye view panoramic view;

[0008] S3, the automatic parking controller analyzes and calculates the bird's-eye view panoramic image, obtains the coordinate information of all parking spaces in the image, and selects the parking space closest to the vehicle as the target parking space;

[0009] S4. Based on the coordinate information of the target parking space, a detection area is formed on the bird's-eye view panoramic image, and brightness detection is performed on the detection area using OpenCV;

[0010] S5. If the brightness does not meet the set detection standard, the target parking space is determined to be in a dark light area, and the fill light is controlled to ensure that the target parking space is always within the set target light intensity range; otherwise, the target parking space is determined not to be in a dark light area, and there is no need to control the fill light;

[0011] S6. The ultrasonic radar monitors the spatial information corresponding to the target parking space in real time. The automatic parking controller integrates the spatial information with the coordinate information of the target parking space and outputs a valid parking space based on the confidence level.

[0012] S7. The automatic parking controller plans a parking path according to the available parking space, and controls the vehicle to complete automatic parking in the available parking space according to the parking path.

[0013] The present invention uses a light sensor and a surround-view camera to synchronously collect the light environment and images around the vehicle in real time. When the ambient light brightness is lower than a certain brightness, the fill light is activated to perform fill light in real time to make the image collected by the surround-view camera clearer. The image collected by the surround-view camera is then analyzed to form a bird's-eye view panoramic image. The coordinate information of the target parking space is obtained according to the bird's-eye view panoramic image, and the brightness of the area where the target parking space is located is detected to determine whether the target parking space is in a dark light area. If so, the fill light is used to compensate the light intensity of the target parking space and adjust the irradiation angle of the fill light to ensure that the target parking space is always in the most suitable brightness environment, thereby achieving the recognition accuracy and recognition success rate of the vehicle to the target parking space in a dark light environment. Finally, the coordinate information of the target parking space is integrated with the parking space space information collected by the ultrasonic radar to complete the parking action.

[0014] Furthermore, in step S2, if the ambient light brightness is less than the set brightness threshold and the duration is greater than the set duration threshold, the fill light is turned on to illuminate the dark area; otherwise, the fill light is not turned on.

[0015] The fill light of the present invention is controlled independently from the surround-view camera. Whether the fill light is turned on does not affect the real-time image collection of the surround-view camera. The fill light is turned on when the light level is lower than the above-mentioned threshold condition to improve the clarity of the image collected by the surround-view camera in a dark environment.

[0016] Furthermore, the brightness threshold is 20 Lux, and the duration threshold is 10 seconds. These two threshold conditions can be set according to actual conditions and are not limited here. They can be changed according to actual needs.

[0017] Furthermore, in step S2, the collected image is processed to obtain a bird's-eye view panoramic image as follows:

[0018] Through the bird's-eye view transformation, multiple distortion-corrected and pre-processed images are pivoted to the same plane according to the camera calibration data, and then a bird's-eye view panoramic view is obtained after image stitching, color balance and brightness consistency processing.

[0019] Furthermore, in step S3, the automatic parking controller analyzes and calculates the bird's-eye view panoramic image to obtain the coordinate information of all parking spaces in the image. The specific process is as follows:

[0020] The edge grayscale image is extracted based on the color information and grayscale gradient information of the bird's-eye view panoramic image, and then the parking space lines are extracted through Hough transform. Then, the corner points are obtained through camera calibration and conversion detection between the pixel coordinate system and the vehicle coordinate system. In this way, the corner points and parking space line coordinates of all parking spaces are obtained.

[0021] Furthermore, in step S5, the detection standard is that the brightness of the detection area is less than 20 Lux for 10 consecutive frames. This detection standard can be set according to actual conditions and is not limited here. It can be changed according to actual needs.

[0022] Furthermore, in step S5, the process of controlling the fill light is as follows:

[0023] Based on the difference between the brightness of the detection area and the detection standard, the fill light's luminous power is adjusted according to a compensation mechanism of increasing 0.5 watts for every 5 Lux reduction. The fill light's light position is also adjusted according to the coordinate information of the target parking space to ensure that the target parking space is always within the set target light intensity range.

[0024] Furthermore, the target illumination intensity range is 80-120 Lux. This range can be set according to actual conditions and is not limited here, and can be changed according to actual needs.

[0025] Furthermore, in step S2, before processing the image, the automatic parking controller first inputs the image captured by the surround-view camera into a preset deep learning neural network model, outputs the optimized image after optimization by the deep learning neural network model, and then processes the optimized image to obtain a bird's-eye view panoramic view; wherein the deep learning neural network model is obtained by training image data collected under various conditions.

[0026] The images captured by the surround-view camera include horizontal parking lines, vertical parking lines, and diagonal parking lines in indoor and outdoor parking lots under different brightness, different colors, different degrees of incompleteness, different lane line interference, different weather conditions, and different site materials. These images are output to a neural network model based on a deep learning algorithm for optimization before output. This can effectively improve the success rate of recognizing parking lines with different brightness, different colors, different degrees of incompleteness, and different lane line interference in subsequent images, thereby improving the accuracy of the coordinate information of the target parking space obtained in subsequent image analysis and calculation.

[0027] The present invention also provides a system for improving parking space recognition rate and parking capability in dark environments, the system is used to implement the above method, and the system includes a light sensor, a fill light, a surround view camera, an ultrasonic radar, and an automatic parking controller;

[0028] The light sensor and fill light are installed on the rearview mirrors on both sides of the vehicle. The light sensor is used to monitor the ambient light brightness around the vehicle, and the fill light is used to provide fill light in a dark environment.

[0029] The surround-view cameras are installed on the four sides of the vehicle and are used to collect images of the vehicle's surroundings;

[0030] The front and rear bumpers of the vehicle are both equipped with ultrasonic radars, which are used to sense environmental information around the vehicle;

[0031] The automatic parking controller is used to control the vehicle to complete the automatic parking action. The automatic parking controller is electrically connected to the light sensor, fill light, surround view camera, and ultrasonic radar respectively.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] This system uses a light sensor, fill light, surround-view camera, ultrasonic radar, and automatic parking controller to sense the ambient light level around the vehicle and capture an image of the parking space. It then analyzes and calculates the coordinates of the target parking space in well-lit conditions. This information is then combined with the spatial information collected by the ultrasonic radar to output a valid parking space. Finally, the automatic parking controller plans a parking route based on the available spaces and completes automatic parking. This system effectively improves the accuracy and success rate of automatic parking space recognition, thereby enhancing the user's parking experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of a method for improving parking space recognition rate and parking capability in a dark environment according to the present invention;

[0035] Figure 2 Schematic diagram of a system for improving parking space recognition and parking capability in low-light environments according to the present invention. DETAILED DESCRIPTION

[0036] The drawings are for illustrative purposes only and should not be construed as limiting this patent. To better illustrate the embodiments, some components in the drawings may be omitted, enlarged, or reduced in size, and do not represent actual product dimensions. Those skilled in the art will understand that some well-known structures and their descriptions may be omitted from the drawings. The positional relationships depicted in the drawings are for illustrative purposes only and should not be construed as limiting this patent.

[0037] Example 1:

[0038] like Figure 1 As shown, this embodiment provides a method for improving parking space recognition rate and parking capability in a dark environment. The specific steps are as follows:

[0039] S1. After the vehicle's automatic parking function is turned on, the vehicle's ambient light brightness is monitored in real time by the light sensor, and the surround-view camera collects images of the vehicle's surroundings in real time;

[0040] S2. The automatic parking controller determines whether to turn on the fill light based on the collected ambient light brightness, and processes the collected image to obtain a bird's-eye view panoramic view;

[0041] S3, the automatic parking controller analyzes and calculates the bird's-eye view panoramic image, obtains the coordinate information of all parking spaces in the image, and selects the parking space closest to the vehicle as the target parking space;

[0042] S4. Based on the coordinate information of the target parking space, a detection area is formed on the bird's-eye view panoramic image, and brightness detection is performed on the detection area using OpenCV;

[0043] S5. If the brightness does not meet the set detection standard, the target parking space is determined to be in a dark light area, and the fill light is controlled to ensure that the target parking space is always within the set target light intensity range; otherwise, the target parking space is determined not to be in a dark light area, and there is no need to control the fill light;

[0044] S6. The ultrasonic radar monitors the spatial information corresponding to the target parking space in real time. The automatic parking controller integrates the spatial information with the coordinate information of the target parking space and outputs a valid parking space based on the confidence level.

[0045] S7. The automatic parking controller plans a parking path according to the available parking space, and controls the vehicle to complete automatic parking in the available parking space according to the parking path.

[0046] The present invention uses a light sensor and a surround-view camera to synchronously collect the light environment and images around the vehicle in real time. When the ambient light brightness is lower than a certain brightness, the fill light is activated to perform fill light in real time to make the image collected by the surround-view camera clearer. The image collected by the surround-view camera is then analyzed to form a bird's-eye view panoramic image. The coordinate information of the target parking space is obtained according to the bird's-eye view panoramic image, and the brightness of the area where the target parking space is located is detected to determine whether the target parking space is in a dark light area. If so, the fill light is used to compensate the light intensity of the target parking space and adjust the irradiation angle of the fill light to ensure that the target parking space is always in the most suitable brightness environment, thereby achieving the recognition accuracy and recognition success rate of the vehicle to the target parking space in a dark light environment. Finally, the coordinate information of the target parking space is integrated with the parking space space information collected by the ultrasonic radar to complete the parking action.

[0047] In step S2 of this embodiment, if the ambient light brightness is less than a set brightness threshold and the duration is greater than a set duration threshold, the fill light is turned on to illuminate the dark area; otherwise, the fill light is not turned on. The brightness threshold is 20 Lux and the duration threshold is 10 seconds.

[0048] The fill light of the present invention is controlled independently from the surround-view camera. Whether the fill light is on or off does not affect the real-time image capture of the surround-view camera. The fill light illuminates when the light level falls below the aforementioned threshold, improving the clarity of the image captured by the surround-view camera in low-light environments. The brightness threshold and duration threshold can be set according to actual conditions and are not limited here. They can be changed according to actual needs.

[0049] In step S2 of this embodiment, the process of processing the collected image to obtain the bird's-eye view panoramic image is as follows:

[0050] Through the bird's-eye view transformation, the four distortion-corrected and pre-processed images are pivoted to the same plane according to the camera calibration data. Then, after image stitching, color balancing and brightness consistency processing, a bird's-eye view panoramic view is obtained.

[0051] In step S3 of this embodiment, the automatic parking controller analyzes and calculates the bird's-eye view panoramic image to obtain the coordinate information of all parking spaces in the image. The specific process is as follows:

[0052] The edge grayscale image is extracted based on the color information and grayscale gradient information of the bird's-eye view panoramic image, and then the parking space lines are extracted through Hough transform. Then, the corner points are obtained through camera calibration and conversion detection between the pixel coordinate system and the vehicle coordinate system. In this way, the corner points and parking space line coordinates of all parking spaces are obtained.

[0053] In this embodiment, the technical means of performing brightness detection on the detection area by using OpenCV is well known to those skilled in the art and will not be described here.

[0054] In step S5 of this embodiment, the detection standard is that the brightness of the detection area is less than 20 Lux for 10 consecutive frames. This detection standard can be set according to actual conditions and is not limited here. It can be changed according to actual needs.

[0055] And in step S5, the process of controlling the fill light is as follows:

[0056] Based on the difference between the brightness of the detection area and the detection standard, the fill light's luminous power is adjusted according to a compensation mechanism of increasing 0.5 watts for every 5 Lux reduction. The fill light's light position is also adjusted according to the coordinate information of the target parking space to ensure that the target parking space is always within the set target light intensity range.

[0057] During specific operation, the automatic parking controller can send a command control to enable the fill light to adjust its light position and light power, wherein the light position corresponds to the dark light area.

[0058] In this embodiment, the target light intensity range can be set to 80-120 Lux. This range can be set according to actual conditions and is not limited here. It can be changed according to actual needs.

[0059] This system uses a light sensor, fill light, surround-view camera, ultrasonic radar, and automatic parking controller to sense the ambient light level around the vehicle and capture an image of the parking space. It then analyzes and calculates the coordinates of the target parking space in well-lit conditions. This information is then combined with the spatial information collected by the ultrasonic radar to output a valid parking space. Finally, the automatic parking controller plans a parking route based on the available spaces and completes automatic parking. This system effectively improves the accuracy and success rate of automatic parking space recognition, thereby enhancing the user's parking experience.

[0060] Example 2:

[0061] like Figure 2 As shown, this embodiment provides a system for improving parking space recognition rate and parking capability in dark environments. The system is used to implement the method in the above embodiment 1. The system includes a light sensor, a fill light, a surround-view camera, an ultrasonic radar, and an automatic parking controller.

[0062] Light sensors and fill lights are installed on the rearview mirrors on both sides of the vehicle. The light sensors are used to monitor the ambient light brightness around the vehicle, and the fill lights are used to provide fill light in dark environments.

[0063] Surround-view cameras are installed on all four sides of the vehicle to collect images of the vehicle's surroundings;

[0064] The front and rear bumpers of the vehicle are equipped with ultrasonic radars, which are used to sense the environmental information around the vehicle;

[0065] The automatic parking controller is used to control the vehicle to complete automatic parking. The automatic parking controller is electrically connected to the light sensor, fill light, surround-view camera, and ultrasonic radar.

[0066] In this embodiment, four surround-view cameras are installed on the vehicle's front grille, left / right rearview mirrors and rear bumper license plate respectively. At the same time, 12 ultrasonic radars are provided, 6 of which are installed on the front bumper and 6 on the rear bumper, so as to achieve all-round perception of the vehicle's surrounding environment information.

[0067] In this embodiment, the surround view camera is a 360° fisheye camera.

[0068] Example 3:

[0069] This embodiment is similar to embodiment 1, except that, in step S2, before processing the image, the automatic parking controller first inputs the image captured by the surround-view camera into a preset deep learning neural network model, and after optimization by the deep learning neural network model, outputs the optimized image, which is then processed to obtain a bird's-eye view panoramic view.

[0070] In this embodiment, the deep learning neural network model is trained by image data collected under various conditions.

[0071] The images captured by the surround-view camera include horizontal parking lines, vertical parking lines, and diagonal parking lines in indoor and outdoor parking lots under different brightness, different colors, different degrees of incompleteness, different lane line interference, different weather conditions, and different site materials. These images are output to a neural network model based on a deep learning algorithm for optimization before output. This can effectively improve the success rate of recognizing parking lines with different brightness, different colors, different degrees of incompleteness, and different lane line interference in subsequent images, thereby improving the accuracy of the coordinate information of the target parking space obtained in subsequent image analysis and calculation.

[0072] Obviously, the above embodiments of the present invention are merely examples for the purpose of clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A method for improving parking space recognition rate and parking capability in dark environments, characterized by: The specific steps are as follows: S1. After the vehicle's automatic parking function is turned on, the vehicle's ambient light brightness is monitored in real time by the light sensor, and the surround-view camera collects images of the vehicle's surroundings in real time; S2. The automatic parking controller determines whether to turn on the fill light based on the collected ambient light brightness, and processes the collected image to obtain a bird's-eye view panoramic view; S3, the automatic parking controller analyzes and calculates the bird's-eye view panoramic image, obtains the coordinate information of all parking spaces in the image, and selects the parking space closest to the vehicle as the target parking space; S4. Based on the coordinate information of the target parking space, a detection area is formed on the bird's-eye view panoramic image, and brightness detection is performed on the detection area using OpenCV; S5. If the brightness does not meet the set detection standard, the target parking space is determined to be in a dark light area, and the fill light is controlled to ensure that the target parking space is always within the set target light intensity range; otherwise, the target parking space is determined not to be in a dark light area, and there is no need to control the fill light; S6. The ultrasonic radar monitors the spatial information corresponding to the target parking space in real time. The automatic parking controller integrates the spatial information with the coordinate information of the target parking space and outputs a valid parking space based on the confidence level. S7. The automatic parking controller plans a parking path according to the available parking space and controls the vehicle to complete automatic parking in the available parking space according to the parking path; In step S2, if the ambient light brightness is less than the set brightness threshold and the duration is greater than the set duration threshold, the fill light is turned on to illuminate the dark area; otherwise, the fill light is not turned on; in step S5, the detection standard is that the brightness of the detection area is less than 20 Lux for 10 consecutive frames; In step S5, the process of controlling the fill light is as follows: Based on the difference between the brightness of the detection area and the detection standard, the fill light's luminous power is adjusted according to a compensation mechanism of increasing 0.5 watts for every 5 Lux reduction. The fill light's light position is also adjusted according to the coordinate information of the target parking space to ensure that the target parking space is always within the set target light intensity range.

2. The method for improving parking space recognition rate and parking capability in low-light environments according to claim 1, characterized in that: The brightness threshold is 20 Lux, and the duration threshold is 10 seconds.

3. The method for improving parking space recognition rate and parking capability in low-light environments according to claim 1, characterized in that: In step S2, the collected image is processed to obtain a bird's-eye view panoramic image as follows: Through the bird's-eye view transformation, multiple distortion-corrected and pre-processed images are pivoted to the same plane according to the camera calibration data, and then a bird's-eye view panoramic view is obtained after image stitching, color balance and brightness consistency processing.

4. The method for improving parking space recognition rate and parking capability in low-light environments according to claim 1, characterized in that: In step S3, the automatic parking controller analyzes and calculates the bird's-eye view panoramic image to obtain the coordinate information of all parking spaces in the image. The specific process is as follows: The edge grayscale image is extracted based on the color information and grayscale gradient information of the bird's-eye view panoramic image, and then the parking space lines are extracted through Hough transform. Then, the corner points are obtained through camera calibration and conversion detection between the pixel coordinate system and the vehicle coordinate system. In this way, the corner points and parking space line coordinates of all parking spaces are obtained.

5. The method for improving parking space recognition rate and parking capability in dark environments according to claim 4, characterized in that: The target light intensity range is 80-120Lux.

6. The method for improving parking space recognition rate and parking capability in low-light environments according to claim 1, characterized in that: In step S2, before processing the image, the automatic parking controller first inputs the image captured by the surround-view camera into a preset deep learning neural network model. After optimization by the deep learning neural network model, the optimized image is output, and then the optimized image is processed to obtain a bird's-eye view panoramic view; wherein the deep learning neural network model is obtained by training image data collected under various conditions.

7. A system for improving parking space recognition and parking capabilities in low-light environments, characterized by: The system is used to implement the method of improving the parking space recognition rate and parking ability in a dark environment as described in any one of claims 1 to 6 above, and the system includes a light sensor, a fill light, a surround-view camera, an ultrasonic radar, and an automatic parking controller; The light sensor and fill light are installed on the rearview mirrors on both sides of the vehicle. The light sensor is used to monitor the ambient light brightness around the vehicle, and the fill light is used to provide fill light in a dark environment. The surround-view cameras are installed on all four sides of the vehicle, and are used to collect images of the vehicle's surroundings; the ultrasonic radars are installed on the front and rear bumpers of the vehicle, and are used to sense environmental information around the vehicle; The automatic parking controller is used to control the vehicle to complete the automatic parking action. The automatic parking controller is electrically connected to the light sensor, fill light, surround view camera, and ultrasonic radar respectively.

Images