Lane Line Detection Method, Device, Equipment and Storage Medium Based on Deep Learning

Abstract

The present invention discloses a lane line detection method, device, equipment and storage medium based on deep learning. The method includes obtaining an original image data set of lane lines and marking the lane lines in the original image data set to obtain a training image data set; wherein, the images in the training image data set include corresponding ground truth boxes of lane lines; inputting the training image data set into a lane line detection model to obtain lane line detection boxes, and completing the training of the lane line detection model according to the loss value between the lane line detection boxes and the ground truth boxes of lane lines; when receiving image data to be detected, inputting the image data to be detected into the lane line detection model to obtain the target lane lines in the image to be detected. The present invention calculates the loss value between the ground truth boxes of lane lines and the lane line detection boxes in the image to train the lane line detection model, improves the correlation between lane line sampling points, and improves the detection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a lane line detection method, device, equipment and storage medium based on deep learning. Background Art

[0002] Lane line detection is an important field of autonomous driving. It refers to the detection and recognition of various lane lines in an image. The application of lane line detection is very extensive, and typical applications are, for example, autonomous driving and assisted driving.

[0003] Existing lane line detection technologies usually include the following: 1. First, use image segmentation technology to obtain a segmentation result, and then perform line fitting on the segmented result. 2. Image detection, sample the image at fixed intervals to obtain many candidate anchor points of lines, and at the same time judge the presence and specific position of lane lines. However, the L1 loss function is often used for line regression, which reduces the association between points and leads to a decrease in the overall effect. Therefore, how to improve the accuracy of lane line detection is a technical problem that needs to be solved urgently.

[0004] The above content is only used to assist in understanding the technical solution of the present invention, and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main object of the present invention is to provide a lane line detection method, device, equipment and storage medium based on deep learning, aiming to solve the technical problem of low accuracy of current lane line detection.

[0006] To achieve the above object, the present invention provides a lane line detection method based on deep learning, and the method includes the following steps:

[0007] Obtain the original image dataset of lane lines, and mark the lane lines in the original image dataset to obtain a training image dataset; wherein, the images in the training image dataset include corresponding ground truth boxes of lane lines;

[0008] Input the training image dataset into a lane line detection model to obtain lane line detection boxes, and complete the training of the lane line detection model according to the loss value between the lane line detection boxes and the ground truth boxes of lane lines;

[0009] When receiving the image data to be detected, input the image data to be detected into the lane line detection model to obtain the target lane lines in the image to be detected.

[0010] Optionally, the step of marking the lane lines in the original image dataset to obtain a training image dataset specifically includes:

[0011] Generate a bounding rectangle for the lane lines in the original image dataset;

[0012] Generate a ground truth box for the lane lines in the original image dataset according to the bounding rectangle.

[0013] Optionally, the size expression of the ground truth box for the lane lines is:

[0014] h = max(h0, w0 / 2);

[0015] w = max(w0, h0 / 2);

[0016] where h0 is the height of the bounding rectangle, w0 is the width of the bounding rectangle, h is the height of the ground truth box for the lane lines, and w is the width of the ground truth box for the lane lines.

[0017] Optionally, the step of inputting the training image dataset into the lane line detection model to obtain the lane line detection box specifically includes:

[0018] Input the training image dataset into the lane line detection model;

[0019] Adjust the positions of the anchor boxes corresponding to each training image in the training image dataset to output the lane line prediction box of the training image.

[0020] Optionally, the step of adjusting the positions of the anchor boxes corresponding to each training image in the training image dataset specifically includes:

[0021] Obtain the predicted class and offset of the anchor boxes pre-determined by the lane line detection model;

[0022] When receiving a training image, adjust the positions of the anchor boxes corresponding to the training image based on the lane line detection model.

[0023] Optionally, before the step of inputting the training image dataset into the lane line detection model, the method further includes:

[0024] Sample the lane lines in the ground truth boxes for the lane lines in the training image dataset to generate lane line sampling points;

[0025] Generate an extended box for the lane line sampling points; wherein, the area on the left side and the right side of the lane line in the extended box is equal;

[0026] Adjust the shape of the ground truth box for the lane lines to a polygon according to the extended box.

[0027] Optionally, the step of adjusting the shape of the ground truth box for the lane lines to a polygon specifically includes:

[0028] Determine whether the point at the lower right corner of the extended box is after the lane line sampling point. If so, use this point and the lane line sampling point to adjust the shape of the true lane line box.

[0029] In addition, to achieve the above object, the present invention also provides a deep learning-based lane line detection device, and the deep learning-based lane line detection device includes:

[0030] An acquisition module, configured to acquire an original image data set of lane lines, and mark the lane lines in the original image data set to obtain a training image data set; wherein, the images in the training image data set include corresponding true lane line boxes;

[0031] A training module, configured to input the training image data set into a lane line detection model to obtain a lane line detection box, and complete the training of the lane line detection model according to the loss value between the lane line detection box and the true lane line box;

[0032] A detection module, configured to input the to-be-detected image data into the lane line detection model to obtain the target lane lines in the to-be-detected image when receiving the to-be-detected image data.

[0033] In addition, to achieve the above object, the present invention also provides a deep learning-based lane line detection device, and the deep learning-based lane line detection device includes: a memory, a processor, and a deep learning-based lane line detection program stored on the memory and executable on the processor. When the deep learning-based lane line detection program is executed by the processor, the steps of the above-mentioned deep learning-based lane line detection method are implemented.

[0034] In addition, to achieve the above object, the present invention also provides a storage medium, on which a deep learning-based lane line detection program is stored. When the deep learning-based lane line detection program is executed by a processor, the steps of the above-mentioned deep learning-based lane line detection method are implemented.

[0035] A lane line detection method, device, equipment and storage medium based on deep learning proposed by an embodiment of the present invention. The method includes obtaining an original image data set of lane lines, marking the lane lines in the original image data set to obtain a training image data set; wherein, the images in the training image data set include corresponding ground truth boxes of lane lines; inputting the training image data set into a lane line detection model to obtain lane line detection boxes, and completing the training of the lane line detection model according to the loss value between the lane line detection boxes and the ground truth boxes of lane lines; when receiving image data to be detected, inputting the image data to be detected into the lane line detection model to obtain the target lane lines in the image to be detected. The present invention calculates the loss value between the ground truth boxes of lane lines and the lane line detection boxes in the image to train the lane line detection model, improves the correlation between lane line sampling points, and improves the detection accuracy. Description of the Drawings

[0036] Figure 1 It is a schematic structural diagram of a lane line detection device based on deep learning in an embodiment of the present invention;

[0037] Figure 2 It is a schematic flowchart of an embodiment of a lane line detection method based on deep learning of the present invention;

[0038] Figure 3 It is a structural block diagram of a lane line detection device based on deep learning in an embodiment of the present invention.

[0039] The realization, functional characteristics and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. Detailed Embodiments

[0040] 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.

[0041] Lane line detection is an important field of autonomous driving. It refers to the detection and recognition of various lane lines in an image. The application of lane line detection is very extensive. A typical application is, for example, autonomous driving and assisted driving.

[0042] The existing lane line detection technologies usually include the following: 1. First, use image segmentation technology to obtain the segmentation result, and then perform line fitting on the segmented result. 2. Image detection, sample the image at fixed intervals to obtain many candidate anchor points of lines, and at the same time judge the presence and specific position of the lane lines. However, the L1 loss function is often used for line regression, which reduces the association between points and leads to a reduction in the overall effect. Therefore, how to improve the accuracy of lane line detection is an urgent technical problem to be solved.

[0043] To solve this problem, various embodiments of the lane line detection method based on deep learning of the present invention are proposed. A lane line detection method, device, equipment and storage medium based on deep learning provided by the present invention, the method includes obtaining an original image data set of lane lines, and marking the lane lines in the original image data set to obtain a training image data set; wherein, the images in the training image data set include corresponding lane line ground truth boxes; inputting the training image data set into a lane line detection model to obtain a lane line detection box, and completing the training of the lane line detection model according to the loss value between the lane line detection box and the lane line ground truth box; when receiving image data to be detected, inputting the image data to be detected into the lane line detection model to obtain the target lane lines in the image to be detected. The present invention calculates the loss value between the lane line ground truth box and the lane line detection box of the image to train the lane line detection model, improves the correlation between lane line sampling points, and improves the detection accuracy.

[0044] Referring to Figure 1 , Figure 1 FIG. is a schematic structural diagram of a lane line detection device based on deep learning according to an embodiment of the present invention.

[0045] The device may be a user equipment (UE) such as a mobile phone, a smart phone, a laptop computer, a digital broadcast receiver, a personal digital assistant (PDA), a tablet computer (PAD), a handheld device, a vehicle-mounted device, a wearable device, a computing device or other processing devices connected to a wireless modem, a mobile station (MS), etc. The device may be referred to as a user terminal, a portable terminal, a desktop terminal, etc.

[0046] Generally, the device includes: at least one processor 301, a memory 302, and a lane line detection program based on deep learning stored on the memory and executable on the processor, and the lane line detection program based on deep learning is configured to implement the steps of the lane line detection method as described above.

[0047] The processor 301 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. The processor 301 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). The processor 301 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the wake state, also known as the CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 301 may be integrated with a GPU (Graphics Processing Unit), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. The processor 301 may also include an AI (Artificial Intelligence) processor, which is used to process the lane line detection operation based on deep learning, so that the lane line detection model based on deep learning can autonomously train and learn to improve efficiency and accuracy.

[0048] The memory 302 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 302 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In some embodiments, the non-transitory computer-readable storage media in the memory 302 is used to store at least one instruction, and the at least one instruction is used to be executed by the processor 801 to implement the lane line detection method based on deep learning provided in the method embodiments of the present application.

[0049] In some embodiments, the terminal may also optionally include: a communication interface 303 and at least one peripheral device. The processor 301, the memory 302, and the communication interface 303 may be connected through a bus or signal lines. Each peripheral device may be connected to the communication interface 303 through a bus, signal lines, or a circuit board. Specifically, the peripheral devices include at least one of a radio frequency circuit 304, a display screen 305, and a power supply 306.

[0050] The communication interface 303 can be used to connect at least one peripheral device related to I / O (Input / Output) to the processor 301 and the memory 302. The communication interface 303 is used by the peripheral device to receive the movement trajectories and other data of multiple mobile terminals uploaded by the user. In some embodiments, the processor 301, the memory 302, and the communication interface 303 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 301, the memory 302, and the communication interface 303 can be implemented on separate chips or circuit boards, and this embodiment does not limit this.

[0051] The radio frequency circuit 304 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The radio frequency circuit 304 communicates with the communication network and other communication devices through electromagnetic signals, so as to obtain the movement trajectories and other data of multiple mobile terminals. The radio frequency circuit 304 converts an electrical signal into an electromagnetic signal for transmission, or converts the received electromagnetic signal into an electrical signal. Optionally, the radio frequency circuit 304 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, and so on. The radio frequency circuit 304 can communicate with other terminals through at least one wireless communication protocol. The wireless communication protocol includes but is not limited to: metropolitan area network, each generation of mobile communication network (2G, 3G, 4G, and 5G), wireless local area network, and / or WiFi (Wireless Fidelity) network. In some embodiments, the radio frequency circuit 304 may also include a circuit related to NFC (Near Field Communication), and this application does not limit this.

[0052] The display screen 305 is used to display the UI (User Interface). The UI may include graphics, text, icons, videos, and any combination thereof. When the display screen 305 is a touch display screen, the display screen 305 also has the ability to collect touch signals on or above the surface of the display screen 305. The touch signal can be input to the processor 301 for processing as a control signal. At this time, the display screen 305 can also be used to provide virtual buttons and / or virtual keyboards, also known as soft buttons and / or soft keyboards. In some embodiments, the display screen 305 can be one, the front panel of the electronic device; in other embodiments, the display screen 305 can be at least two, respectively arranged on different surfaces of the electronic device or in a foldable design; in still other embodiments, the display screen 305 can be a flexible display screen, arranged on the curved surface or the folding surface of the electronic device. Even, the display screen 305 can be set to an irregular non-rectangular shape, that is, a special-shaped screen. The display screen 305 can be prepared from materials such as LCD (Liquid Crystal Display) and OLED (Organic Light-Emitting Diode).

[0053] The power supply 306 is used to supply power to each component in the electronic device. The power supply 306 can be alternating current, direct current, a disposable battery, or a rechargeable battery. When the power supply 306 includes a rechargeable battery, the rechargeable battery can support wired charging or wireless charging. The rechargeable battery can also be used to support fast charging technology.

[0054] Those skilled in the art can understand that Figure 1 the structure shown in does not constitute a limitation on the lane line detection device based on deep learning, and may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0055] An embodiment of the present invention provides a lane line detection method based on deep learning. Refer to Figure 2 , Figure 2 which is a schematic flowchart of the first embodiment of the lane line detection method based on deep learning of the present invention.

[0056] In this embodiment, the lane line detection method based on deep learning includes the following steps:

[0057] Step S100, obtain the original image dataset of the lane line, and mark the lane lines in the original image dataset to obtain the training image dataset; wherein, the images in the training image dataset contain the corresponding lane line ground truth boxes.

[0058] Specifically, after obtaining the original image data of the lane line, mark the lane line in the original image data to obtain the training image data with the true box of the lane line.

[0059] It should be noted that in this embodiment, by marking the lane line in the original image data set, a training image data set is obtained. By generating a bounding rectangle for the lane line in the original image data set, and based on the bounding rectangle, generating a true box for the lane line in the original image data set. Furthermore, the training of the lane line detection model can be carried out according to the true box and the subsequent detection box.

[0060] It is easy to understand that the size expression of the true box of the lane line is:

[0061] h = max(h0, w0 / 2);

[0062] w = max(w0, h0 / 2);

[0063] Where h0 is the height of the bounding rectangle, w0 is the width of the bounding rectangle, h is the height of the true box of the lane line, and w is the width of the true box of the lane line.

[0064] Furthermore, after generating the training image data with the true box of the lane line, sampling can be carried out on the lane line in the true box of the lane line in the training image data set to generate lane line sampling points; generating an extended box for the lane line sampling points; wherein, the area on the left side and the right side of the lane line in the extended box is equal; according to the extended box, adjusting the shape of the true box of the lane line to a polygon.

[0065] It is easy to understand that the extended box is a square with both width and height equal to the length of the lane line, and this extended box contains the vertically sampled lane line points. Therefore, adjusting the shape of the true box of the lane line to a polygon can be achieved by judging whether the point at the lower right corner of the extended box is after the sampling point. If so, include the point at the lower right corner into the true box of the lane line, and thus connect multiple points to form a polygon.

[0066] Step S200, input the training image data set into the lane line detection model to obtain the lane line detection box, and complete the training of the lane line detection model according to the loss value between the lane line detection box and the true box of the lane line.

[0067] Specifically, after obtaining the training image data set, input the training image data set into the lane line detection model. Before this, it is necessary to first obtain the predicted category and offset of the anchor box pre-determined by the lane line detection model. Furthermore, when receiving the training image, based on the lane line detection model, adjust the position of the anchor box corresponding to the training image.

[0068] Further, after the lane line detection model receives the training image dataset, adjust the positions of the anchor boxes corresponding to each training image in the training image dataset to output the lane line prediction boxes for the training image.

[0069] Finally, use the loss values of the ground truth boxes and the lane line prediction boxes of each training image to train the lane line detection model until the model converges, completing the training process.

[0070] Step S300, when receiving the image data to be detected, input the image data to be detected into the lane line detection model to obtain the target lane lines in the image to be detected.

[0071] Specifically, after obtaining the trained lane line detection model, wait for the input of the image data to be detected. If the image data to be detected is received, input the image data to be detected into the trained lane line detection model to obtain the output target lane lines.

[0072] In this embodiment, a lane line detection method based on deep learning is provided. By obtaining the original image dataset of the lane lines and marking the lane lines in the original image dataset, a training image dataset is obtained; wherein, the images in the training image dataset include corresponding ground truth boxes of the lane lines; input the training image dataset into the lane line detection model to obtain the lane line detection boxes, and complete the training of the lane line detection model according to the loss values between the lane line detection boxes and the ground truth boxes of the lane lines; when receiving the image data to be detected, input the image data to be detected into the lane line detection model to obtain the target lane lines in the image to be detected. The present invention calculates the loss values between the ground truth boxes and the lane line detection boxes of the images to train the lane line detection model, improving the correlation between the lane line sampling points and the detection accuracy.

[0073] Refer to Figure 3 , Figure 3 which is the structural block diagram of the embodiment of the lane line detection device based on deep learning of the present invention.

[0074] As Figure 3 shown, the lane line detection device based on deep learning proposed in the embodiment of the present invention includes:

[0075] An acquisition module 10, configured to acquire the original image dataset of the lane lines and mark the lane lines in the original image dataset to obtain a training image dataset; wherein, the images in the training image dataset include corresponding ground truth boxes of the lane lines;

[0076] The training module 20 is configured to input the training image dataset into the lane detection model to obtain lane detection boxes, and complete the training of the lane detection model according to the loss value between the lane detection boxes and the ground truth boxes of the lanes.

[0077] The detection module 30 is configured to input the image data to be detected into the lane detection model to obtain the target lanes in the image data to be detected when receiving the image data to be detected.

[0078] As an implementation manner, the acquisition module 10 is further configured to generate a bounding rectangle for the lanes in the original image dataset; and generate ground truth boxes for the lanes in the original image dataset according to the bounding rectangle.

[0079] As an implementation manner, the size expression of the ground truth box of the lanes in the acquisition module 10 is:

[0080] h = max(h0, w0 / 2);

[0081] w = max(w0, h0 / 2);

[0082] where h0 is the height of the bounding rectangle, w0 is the width of the bounding rectangle, h is the height of the ground truth box of the lane, and w is the width of the ground truth box of the lane.

[0083] As an implementation manner, the training module 20 is further configured to input the training image dataset into the lane detection model; and adjust the positions of the anchor boxes corresponding to each training image in the training image dataset to output the lane prediction boxes of the training image.

[0084] As an implementation manner, the training module 20 is further configured to obtain the predicted categories and offsets of the anchor boxes predetermined by the lane detection model; and adjust the positions of the anchor boxes corresponding to the training image based on the lane detection model when receiving the training image.

[0085] As an implementation manner, the lane detection device based on deep learning further includes an extension module 40. The extension module 40 is further configured to sample the lanes in the ground truth boxes of the lanes in the training image dataset to generate lane sampling points; generate extension boxes for the lane sampling points; wherein, the areas on the left and right sides of the lane in the extension box are equal; and adjust the shape of the ground truth box of the lane to a polygon according to the extension box.

[0086] As an implementation manner, the expansion module 40 is further configured to, when receiving a test image data set, determine whether the stop movement time of the target vehicle in the test image data set exceeds a preset value. If so, determine the target body frame image; and use the chassis detection model to detect the chassis position in the target body frame image to obtain the vertex position information of the chassis in the target body frame image.

[0087] The lane line detection device based on deep learning provided in this embodiment obtains the original image data set of the lane line and marks the lane line in the original image data set to obtain a training image data set. Among them, the images in the training image data set include corresponding lane line ground truth boxes. Input the training image data set into the lane line detection model to obtain a lane line detection box, and complete the training of the lane line detection model according to the loss value between the lane line detection box and the lane line ground truth box. When receiving the image data to be detected, input the image data to be detected into the lane line detection model to obtain the target lane line in the image to be detected. The present invention calculates the loss value between the lane line ground truth box and the lane line detection box of the image to train the lane line detection model, improves the correlation between the lane line sampling points, and improves the detection accuracy.

[0088] For other embodiments or specific implementation manners of the lane line detection device based on deep learning of the present invention, reference may be made to the above method embodiments, and details will not be described herein again.

[0089] In addition, an embodiment of the present invention further provides a storage medium, on which a lane line detection program based on deep learning is stored. When the lane line detection program based on deep learning is executed by a processor, the steps of the lane line detection method based on deep learning as described above are implemented. Therefore, details will not be described herein again. In addition, the beneficial effects of using the same method will not be described again. For the technical details not disclosed in the embodiment of the computer-readable storage medium involved in the present application, please refer to the description of the method embodiment of the present application. By way of example, the program instructions can be deployed to be executed on one computing device, or on multiple computing devices located at one location, or on multiple computing devices distributed at multiple locations and interconnected by a communication network.

[0090] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The above program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above method embodiments. Among them, the above storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0091] In addition, it should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the attached drawings of the device embodiments provided by the present invention, the connection relationships between the modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement this without creative efforts.

[0092] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware. Of course, it can also be implemented by dedicated hardware including application-specific integrated circuits, dedicated CPUs, dedicated memories, dedicated components, etc. Generally, functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structures for implementing the same function can also be various, such as analog circuits, digital circuits or dedicated circuits. However, for the present invention, software program implementation is a better implementation method in more cases. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a floppy disk, USB flash drive, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disc of a computer, and includes several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

Claims

1. A lane line detection method based on deep learning, characterized in that, The method includes the following steps: Obtain the original image dataset of lane lines, and label the lane lines in the original image dataset to obtain a training image dataset; Among them, the images in the training image dataset include corresponding ground truth boxes of lane lines; Specifically, it includes: Sample the lane lines in the ground truth boxes of lane lines in the training image dataset to generate lane line sampling points; Generate an extended box for the lane line sampling points; wherein, the area on the left side and the area on the right side of the lane line in the extended box are equal; According to the extended box, adjust the shape of the ground truth box of the lane line to a polygon; Input the training image dataset into the lane line detection model, adjust the positions of the anchor boxes corresponding to each training image in the training image dataset to output the lane line detection box of the training image; and complete the training of the lane line detection model according to the loss value between the lane line detection box and the ground truth box of the lane line; When receiving the image data to be detected, input the image data to be detected into the lane line detection model to obtain the target lane line in the image to be detected.

2. The lane line detection method based on deep learning according to claim 1, characterized in that, The size expression of the ground truth box of the lane line is: h = max(h0, w0 / 2); w = max(w0, h0 / 2); Among them, h0 is the height of the circumscribed rectangle box, w0 is the width of the circumscribed rectangle box, h is the height of the ground truth box of the lane line, and w is the width of the ground truth box of the lane line.

3. The lane line detection method based on deep learning according to claim 1, wherein The step of adjusting the positions of the anchor boxes corresponding to each training image in the training image dataset specifically includes: Obtain the predicted category and offset of the anchor box pre-determined by the lane line detection model; When receiving a training image, based on the lane line detection model, adjust the position of the anchor box corresponding to the training image.

4. A lane line detection device based on deep learning, characterized in that, The lane line detection device based on deep learning includes: An acquisition module, configured to obtain the original image dataset of lane lines, and label the lane lines in the original image dataset to obtain a training image dataset; wherein, the images in the training image dataset include corresponding ground truth boxes of lane lines; A training module, configured to sample the lane lines in the ground truth boxes of lane lines in the training image dataset to generate lane line sampling points; generate an extended box for the lane line sampling points; wherein, the area on the left side and the area on the right side of the lane line in the extended box are equal; according to the extended box, adjust the shape of the ground truth box of the lane line to a polygon; input the training image dataset into the lane line detection model, adjust the positions of the anchor boxes corresponding to each training image in the training image dataset to output the lane line detection box of the training image; and complete the training of the lane line detection model according to the loss value between the lane line detection box and the ground truth box of the lane line; A detection module, configured to input the image data to be detected into the lane line detection model when receiving the image data to be detected, to obtain the target lane line in the image to be detected.

5. A lane line detection device based on deep learning, characterized in that, The lane line detection device based on deep learning includes: a memory, a processor, and a lane line detection program based on deep learning stored on the memory and executable on the processor. When the lane line detection program based on deep learning is executed by the processor, the steps of the lane line detection method based on deep learning according to any one of claims 1 to 3 are implemented.

6. A storage medium, characterized in that, A lane line detection program based on deep learning is stored on the storage medium. When the lane line detection program based on deep learning is executed by a processor, the steps of the lane line detection method based on deep learning according to any one of claims 1 to 3 are implemented.

Images