Parking drivable area detection method based on cascade forest and stacked ensemble learning

Abstract

The invention belongs to the technical field of computer vision, and provides a parking drivable area detection method based on cascade forest and stacked ensemble learning, and the method comprises the following steps: S1, obtaining a feature matrix of a road image; s2, the feature matrix is input into a cascade forest model, each layer in the cascade forest model comprises a fixed number of multiple tree structure base learners, and the cascade forest model outputs multiple vehicle drivable area segmentation results corresponding to the multiple tree structure base learners; and S3, learning the output of the cascade forest model by using a stacked ensemble learning algorithm to obtain a final vehicle driving region segmentation result. The method provided by the invention solves the problems that the detection accuracy of the drivable area is low and a large number of computing resources are depended in an intelligent parking environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer vision, in particular to a parking drivable area detection method based on cascade forest and stacked ensemble learning. BACKGROUND

[0002] In the parking assistance system, the drivable area refers to the area in which the vehicle can safely drive in the parking environment. Through the vehicle-mounted camera, the system collects the visual information of the parking lot or the road, obtains the key information such as lane lines, ground markings, parking space contours and obstacles, etc. Then, a deep learning model is established by using computer vision technology, the model further analyzes the image, and segments the drivable area. Finally, based on the detected drivable area, the system plans the path to ensure that the vehicle can avoid obstacles and smoothly and safely drive into the target parking space.

[0003] The deep learning model divides the two-dimensional fisheye image input by the vehicle-mounted camera into a drivable area and a background through pixel-by-pixel classification. The model automatically extracts visual features related to the drivable area by learning a large amount of labeled data, and accurately segments the area in which the vehicle can safely drive. However, the features extracted from the image by the deep learning model are usually high-dimensional and complex, and these features are not always optimized for a specific task, requiring a large amount of data and training time to obtain suitable features, which reduces the efficiency and accuracy of drivable area detection in the parking perception process.

[0004] Therefore, the present application provides a parking drivable area detection method based on cascade forest and stacked ensemble learning, which uses tree structure instead of neurons in deep learning as a basic component, and retains the hierarchical architecture in deep learning. The features such as color, texture and edge of the image are extracted from the intelligent parking image, these data with certain prior knowledge are input to the model for classification, then the secondary learning of the classification results is carried out by using the stacked ensemble learning method, and finally the segmented drivable area is obtained. SUMMARY

[0005] The present application aims to solve at least one of the technical problems existing in the prior art, and provides a parking drivable area detection method based on cascade forest and stacked ensemble learning.

[0006] To solve the above technical problems, the first aspect of the present application provides a parking drivable area detection method based on cascade forest and stacked ensemble learning, comprising the following steps:

[0007] S1. Obtain the feature matrix of the road image;

[0008] S2. Inputting the feature matrix into a cascade forest model, wherein each layer of the cascade forest model includes a fixed number of multiple tree-structured base learners, and the cascade forest model outputs multiple vehicle drivable area segmentation results corresponding to the multiple tree-structured base learners;

[0009] S3. Use the stacked ensemble learning algorithm to learn the output of the cascade forest model to obtain the final vehicle drivable area segmentation result.

[0010] Further, step S1 includes:

[0011] performing feature extraction on the road image, wherein the extracted features include at least one of color features, texture features, edge features, and contour features;

[0012] Build based on feature extraction results dimensional feature matrix, where Represents the number of pixel samples in the road image, The dimension of image features representing pixel samples.

[0013] Furthermore, the extraction of the color feature includes the following steps:

[0014] I. Color histogram feature extraction;

[0015] Count the red, green and blue channel values ​​of each pixel in the road image and distribute them in a histogram. By calculating the frequency of occurrence of pixel values ​​in each channel, we can obtain information about the color distribution of the road image.

[0016] II. Color space feature extraction;

[0017] Convert the road image from RGB to hue, saturation, and value color space and extract HSV information. The calculation formula is as follows:

[0018]

[0019]

[0020]

[0021] Where, 、 、 Represents the red, green, and blue channel values ​​of the pixel respectively; Indicates hue; Indicates saturation; Indicates brightness.

[0022] Furthermore, the extraction of the texture features comprises the following steps:

[0023] I. Gray Level Co-occurrence Matrix (GLCM) feature extraction;

[0024] By counting the co-occurrence frequency of grayscale values ​​between pixel pairs in the image, the texture information of the vehicle's drivable area is extracted. The calculation formula is as follows:

[0025]

[0026] Where, is the gray level in the image; According to the angle and distance Determined horizontal displacement; According to the angle and distance Determined vertical displacement; is the size of the image;

[0027] II. Local binary pattern (LBP) feature extraction;

[0028] By comparing the grayscale values ​​of pixels with those of corresponding neighboring pixels, the texture information of the vehicle drivable area in the neighborhood of the pixel point is extracted. The calculation formula is as follows:

[0029]

[0030] in, are the coordinates of the center pixel; The coordinates are The number of neighboring pixels of the central pixel, It is The grayscale value of pixels in the area, and is an integer; is the gray value of the center pixel; function Defined as:

[0031] .

[0032] Furthermore, the extraction of edge features includes the following steps: identifying discontinuities in pixel values ​​using a Canny edge detection algorithm to determine edge positions;

[0033] The calculation formula is as follows:

[0034]

[0035] Where, is the magnitude of the gradient; is the horizontal component of the gradient; is the longitudinal component of the gradient; Characterize the direction of the gradient;

[0036] The local maximum gradient value is retained for the calculated gradient, and the edge is determined using the double threshold method. Finally, the edge pixels are connected to obtain the edge features of the vehicle's drivable area.

[0037] Furthermore, the extraction of the contour features comprises the following steps:

[0038] Describing the contour features in the image by shape descriptors, thereby extracting the geometric features of the object to represent the shape of the object; the shape descriptors include Hu moments and Zernike moments;

[0039] The Hu moment is calculated as follows:

[0040]

[0041]

[0042] Where, is the original moment of the image, is the order; is the coordinate in the image The pixel value at ; is the central moment of the image; are the coordinates of the image's centroid, , ;

[0043]

[0044]

[0045]

[0046] The calculation formula of Zernike moment is as follows:

[0047]

[0048] Where, yes Second-rate Zernike moments of order, is the radial order, is the angular order, satisfying and is an even number; are Zernike orthogonal polynomials.

[0049] Furthermore, in step S3, a support vector machine is used as a meta-learner to learn the output of the cascade forest model.

[0050] Furthermore, the multiple tree structure-based learners include: 2 random forest learners and 2 XGBoost learners.

[0051] To solve the above technical problems, the second aspect of the present application provides an electronic device, comprising:

[0052] one or more processors;

[0053] a memory for storing one or more programs;

[0054] When the one or more programs are executed by the one or more processors, the one or more processors implement the parking drivable area detection method based on cascade forest and stacked ensemble learning as provided in the first aspect above.

[0055] To solve the above technical problems, the third aspect of the present application provides a computer readable medium, wherein the computer readable medium stores a computer program, and the computer program is executed by a processor to implement the parking drivable area detection method based on cascade forest and stacked ensemble learning as provided in the first aspect above.

[0056] The present application has the beneficial effects that: the present application uses cascade forest and stacked ensemble learning method to realize intelligent parking drivable area detection. The present application extracts features such as color, texture and edge of the image from the intelligent parking image, inputs these data with certain prior knowledge into the cascade forest model for classification, solves the problem of low accuracy of drivable area detection in intelligent parking environment. Then the secondary learning of classification result is carried out through the stacked ensemble learning method, and finally the segmented drivable area is obtained, which solves the problem that the drivable area detection in intelligent parking environment depends on a large amount of computing resources. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 A flow chart of a parking drivable area detection method based on cascade forest and stacked ensemble learning provided by the embodiment of the present application;

[0058] Figure 2 A step schematic diagram of a parking drivable area detection method based on cascade forest and stacked ensemble learning provided by the embodiment of the present application;

[0059] Figure 3 A feature extraction diagram of step S1 in the embodiment of the present application;

[0060] Figure 4 A structural block diagram of the electronic device provided by the embodiment of the present application;

[0061] Figure 5 A structural schematic diagram of the computer readable medium provided by the embodiment of the present application. DETAILED DESCRIPTION

[0062] For a better understanding of the technical solutions of the present application, the exemplary embodiments of the present application are described below in conjunction with the drawings, which include various details of the embodiments of the present application to help understanding, and should be considered as merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Also, for the sake of clarity and conciseness, the description below omits the description of well-known functions and structures.

[0063] In the case of no conflict, each embodiment of the present application and each feature in the embodiments can be combined with each other.

[0064] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0065] The terms used herein are only used to describe specific embodiments, and are not intended to limit the present application. As used herein, the singular forms "a" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that the terms "comprise" and / or "consist of", when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. The terms "connected" or "coupled" and / or similar terms are not limited to a physical or mechanical connection, but can include an electrical connection, whether direct or indirect.

[0066] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and the present application, and will not be interpreted in an overly formal or overly literal sense unless expressly so defined herein.

[0067] In the technical solutions of the present application, the collection, storage, use, processing, transmission, provision and disclosure of user personal information comply with relevant laws and regulations, and do not violate public order and good customs. The use of user data in the technical solutions complies with relevant national laws and regulations (for example, "Information Security Technology Personal Information Security Specification" and the like). For example, appropriate measures are taken for personal information access control; restrictions are given for the display of personal information; the use purpose of personal information does not exceed the direct or reasonably related range; the use of personal information eliminates the explicit identity pointing and avoids precise positioning to a specific individual.

[0068] Prior art related to the present application: CN117876396A discloses a drivable area segmentation method and device based on splittable convolution large-scale selective attention, comprising: acquiring real-time images of a road; constructing a drivable area segmentation model based on an improved Deeplabv3+ network and training to obtain a trained drivable area segmentation model, the drivable area segmentation model based on the improved Deeplabv3+ network sets a splittable convolution large-scale selective attention module between the backbone network and the ASPP module in the encoder of the traditional Deeplabv3+ network, the splittable convolution large-scale selective attention module includes a series of large-scale convolution kernel sequences, a spatial kernel selection unit and a first convolution layer; input the real-time image into the trained drivable area segmentation model to obtain the output features of the splittable convolution large-scale selective attention module and input them into the ASPP module, then decode them through the decoder to obtain the segmentation result, which can improve the overall detection accuracy and edge segmentation accuracy.

[0069] However, the prior art has the following disadvantages: 1. The real-time image features of the road do not consider the prior knowledge of the parking perception environment, such as color, texture and edge features, resulting in reduced drivable area recognition accuracy; 2. The model has a large number of parameters, requires a large amount of computing resources, and has slow training and inference speed, especially in small-scale data and resource-limited environments.

[0070] To solve the above problems, the present application provides an intelligent parking drivable area detection method based on cascading forest and stacked ensemble learning, which uses tree structure instead of neurons in deep learning as basic components and retains the hierarchical architecture in deep learning. The color, texture and edge features of the image are extracted from the intelligent parking image, and these data with certain prior knowledge are input into the model for classification, and then the stacked ensemble learning method is used for secondary learning of the classification results, and finally the segmented drivable area is obtained. Through the above method, the problem of low drivable area detection accuracy in intelligent parking environment and dependence on a large amount of computing resources is solved.

[0071] To solve at least one of the technical problems existing in the related art, the present application provides a parking drivable area detection method based on cascading forest and stacked ensemble learning. Figure 1 A flowchart of a parking drivable area detection method based on cascading forest and stacked ensemble learning provided by an embodiment of the present application is provided. Figure 2 A step schematic diagram of a parking drivable area detection method based on cascading forest and stacked ensemble learning provided by an embodiment of the present application is provided. The present embodiment includes the following steps:

[0072] S1. Obtain the feature matrix of the road image;

[0073] The road image is captured in real time by a vehicle-mounted fisheye camera, and features are extracted from the collected road image, wherein the extracted features include at least one of color features, texture features, edge features, and contour features. The extracted features are as shown in Figure 3

[0074] A dimension feature matrix is constructed according to the feature extraction result, wherein represents the number of pixel samples in the road image, represents the image feature dimension of the pixel sample.

[0075] The extraction of color features includes the following steps:

[0076] I. Color histogram feature extraction;

[0077] The red, green, and blue (RGB) channel values of each pixel in the road image are counted and statistically distributed into a histogram. The frequency of the pixel values of each channel is calculated to obtain the color distribution information of the road image.

[0078] II. Color space feature extraction;

[0079] The road image is converted from RGB to hue, saturation, and lightness (Hue Saturation Value, HSV) color space, and HSV information is extracted. The calculation formula is as follows:

[0080]

[0081]

[0082]

[0083] In the formula, , , respectively represent the red, green, and blue channel values of the pixel; represents the hue; represents the saturation; represents the lightness.

[0084] The extraction of texture features includes the following steps:

[0085] I. Gray-Level Co-occurrence Matrix (GLCM) feature extraction;

[0086] ​GLCM is a method for describing the texture features of the vehicle drivable area, which calculates the co-occurrence relationship between pairs of pixels of gray levels. By counting the co-occurrence frequency of gray values between pairs of pixels in the image, the texture information of the vehicle drivable area is extracted, and the calculation formula is as follows:

[0087]

[0088] In the formula, is the gray level in the image; is the horizontal displacement determined according to the angle and the distance ; is the vertical displacement determined according to the angle and the distance ; is the size of the image;

[0089] II. Local Binary Patterns (LBP) feature extraction;

[0090] LBP is used to describe the texture features of the vehicle drivable area in the neighborhood of a pixel, which extracts features by comparing the gray values of the pixel and its neighborhood pixels, and the calculation formula is as follows:

[0091]

[0092] wherein, is the coordinate of the center pixel; is the number of neighborhood pixels of the center pixel with coordinate , is the gray value of the th neighborhood pixel, and is an integer; is the gray value of the center pixel; the function is defined as:

[0093] .

[0094] The extraction of edge features includes the following steps: determining the edge position by identifying the discontinuity of pixel values through the Canny edge detection algorithm;

[0095] The calculation formula is as follows:

[0096]

[0097] In the formula, is the amplitude of the gradient; is the horizontal component of the gradient; is the vertical component of the gradient; represents the direction of the gradient;

[0098] The local maximum gradient value is reserved for the calculated gradient, the edge is determined by using a double threshold method, and finally the edge pixels are connected, so as to obtain the edge features of the vehicle drivable area.

[0099] The extraction of the contour feature includes the following steps: describing the contour feature in the image by a shape descriptor, so as to extract the geometric feature of the object to represent the shape of the object; the shape descriptor includes Hu moment and Zernike moment, and the Hu moment and the Zernike moment are commonly used shape descriptors for representing the shape of the object;

[0100] The calculation method of the Hu moment is as follows:

[0101]

[0102]

[0103] In the formula, is the original moment of the image, is the order; is the pixel value at the coordinate in the image; is the central moment of the image; is the centroid coordinate of the image, , ; and

[0104]

[0105]

[0106]

[0107] The image moment is mainly used for target detection, shape analysis and position estimation in the intelligent parking system of the fish-eye camera, and the specific scenarios are as follows:

[0108] 1. Target centroid calculation (such as parking line and obstacle)

[0109] For the target area (such as lane line and parking space contour) after binarization, the zero-order moment and the first-order moment , are calculated, so as to obtain the centroid .

[0110] Application: The centroid position is used to determine the center of the target (such as a parking space), and to assist in planning the parking path.

[0111] 2. Shape description and classification

[0112] The second-order central moment , , Can be combined into normalized central moments (such as Hu moments) for describing the invariant features of shape (rotation, translation, and scale invariant).

[0113] Application: distinguish the outline of parking space (rectangle) and obstacle (irregular shape), improve the recognition robustness.

[0114] 3. Fish-eye image distortion correction auxiliary

[0115] The actual centroid and direction of known shapes (such as calibration board, parking line) in the scene are calculated using image moment, and compared with the fish-eye distorted image to optimize the distortion correction parameters.

[0116] Zernike moment is a shape description method based on orthogonal Zernike polynomial, and the calculation formula is as follows:

[0117]

[0118] In the formula, is the nth order Zernike moment, is the radial order, is the angular order, and satisfies and is even; is the Zernike orthogonal polynomial;

[0119] In the vehicle-mounted fish-eye vision system, the rotation invariance of Zernike moment can be used to extract the shape features of parking space and obstacle (such as the stable Zernike moment features of rectangular parking space), to assist drivable area segmentation and target recognition, and to adapt to the image analysis requirements of different angles of vehicle shooting.

[0120] S2. Input the feature matrix obtained in step S1 into the cascade forest model, each layer of the cascade forest model includes a fixed number of tree structure base learners, and the cascade forest model outputs a plurality of vehicle drivable area segmentation results corresponding to the plurality of tree structure base learners;

[0121] The main structure of the cascade forest model is an ensemble learning model composed of layers and cascades. In the cascade structure of the cascade forest model, each layer contains a plurality of base learners, and the output of the previous layer and the splicing of the original input features are used as the input of the next layer. In each layer of the cascade forest, there are 2 random forests and 2 XGBoost learners. The learners obtain the input road image features and output the vehicle drivable area segmentation results.

[0122] S3. Use the stacked ensemble learning algorithm to re-learn the output of the cascade forest to obtain the final vehicle drivable area segmentation result;

[0123] Assume that the total number of spatial prediction sample training sets is The total number of validation sets is The total number of test sets is Assume that the input parking perception fisheye image feature set is:

[0124]

[0125] The feature sample training set of the input parking perception fisheye image is:

[0126]

[0127] The test set is:

[0128]

[0129] The validation set is:

[0130]

[0131] During the model training process, the training set and validation set are input into each base learner of the cascade forest model, and k-fold cross validation is performed. The training set is used to fit each base learner step by step, and the validation set is input into the fitted base learner. The corresponding secondary features are extracted from the prediction results of the base learner as follows:

[0132]

[0133] In the stacked ensemble learning method, in order to avoid overfitting, a model with a simpler structure is selected as the meta-learner. In the embodiment of the present invention, the support vector machine is selected as the meta-learner. , input into the meta-learner for fitting, and obtain the final drivable area segmentation result.

[0134] Based on the same inventive concept, an embodiment of the present invention further provides an electronic device. Figure 4 This is a structural block diagram of an electronic device provided by an embodiment of the present invention. Figure 4 As shown, an embodiment of the present invention provides an electronic device comprising: one or more processors 101, a memory 102, and one or more I / O interfaces 103. The memory 102 stores one or more programs, which, when executed by the one or more processors, enable the one or more processors to implement any of the parking drivable area detection methods based on cascade forest and stacked ensemble learning in the above-described embodiments. The one or more I / O interfaces 103 are connected between the processors and the memory and are configured to enable information exchange between the processors and the memory.

[0135] The processor 101 is a device with data processing capability, including but not limited to a central processing unit (CPU) and the like; the memory 102 is a device with data storage capability, including but not limited to a random access memory (RAM, more specifically SDRAM, DDR, etc.), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), and a flash memory (FLASH); and the I / O interface 103 is connected between the processor 101 and the memory 102, and can realize information interaction between the processor 101 and the memory 102, including but not limited to a data bus and the like.

[0136] In some embodiments, the processor 101, the memory 102, and the I / O interface 103 are connected to each other through the bus 104, and further connected to other components of the computing device.

[0137] In some embodiments, the one or more processors 101 include a field programmable gate array.

[0138] The embodiments of the present application also provide a computer readable medium. The computer readable medium stores a computer program, wherein the program is executed by a processor to implement the steps of the above-mentioned parking drivable area detection method based on cascade forest and stacking ensemble learning.

[0139] Figure 5 An embodiment of the computer readable storage medium of the present application is a structural schematic block diagram. The computer readable storage medium 200 stores program data 210, and the program data 210 is executed to implement the steps of the above-mentioned camera and laser radar extrinsic parameter verification method.

[0140] The embodiments of the present application also provide a computer program product, including computer readable code or a non-volatile computer readable storage medium carrying the computer readable code. When the computer readable code is run in the processor of the electronic device, the processor in the electronic device executes the above-mentioned parking drivable area detection method based on cascade forest and stacking ensemble learning.

[0141] Those of ordinary skill in the art will realize and understand that all or some of the steps in the methods disclosed above and the functional modules / units in the systems and devices can be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementation, the division between the functional modules / units referred to in the above description does not necessarily correspond to the division of physical components; for example, one physical component can have multiple functions, or one function or step can be performed by several physical components in cooperation. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software can be distributed on computer readable storage media, which can include computer storage media (or non-transitory media) and communication media (or transitory media).

[0142] As is well known to those of ordinary skill in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable program instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM), static random access memory (SRAM), flash memory or other memory technology, portable compact disc read only memory (CD-ROM), digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be accessed by a computer. Further, it is well known to those of ordinary skill in the art that communication media typically embodies computer readable program instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. As used herein, the term "modulated data signal" means a signal that has one or more of its characteristics changed or set in a manner so as to encode information in the signal. By way of example, and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as wireless networks, cellular telephone networks, code division multiple access (CDMA) networks, and other terrestrial and satellite radio frequency communication networks or frequency (RF) media. The computer software modules and program modules described herein can include routines, programs, functions, objects, components, data structures, and the like that perform particular tasks. The software modules and program modules can be written in any of various programming languages, such as Java™, C, C++, C#, or the like, and can be stored in any computer-readable medium, such as a computer memory, a magnetic disk, or other storage device.

[0143] The computer software modules and program modules described herein can include routines, programs, functions, objects, components, data structures, and the like that perform particular tasks. The software modules and program modules can be written in any of various programming languages, such as Java™, C, C++, C#, or the like, and can be stored in any computer-readable medium, such as a computer memory, a magnetic disk, or other storage device.

[0144] Computer readable program instructions for carrying out operations of the present application can be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate array (FPGA), or programmable logic array (PLA) can execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present application.

[0145] The computer program product described herein can be embodied in a specific manner by hardware, software, or a combination thereof. In an optional embodiment, the computer program product is embodied as a computer storage medium. In another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (SDK), and the like.

[0146] The computer program product described herein can be embodied in a specific manner by hardware, software, or a combination thereof. In an optional embodiment, the computer program product is embodied as a computer storage medium. In another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (SDK), and the like.

[0147] These computer readable program instructions can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer readable program instructions can also be stored in a computer readable storage medium that can include a non-transitory computer readable storage medium that can be a computer- readable storage medium having no data storage cycles that change state. The instructions can be executed by one or more processors of a computer, other programmable data processing apparatus, or other devices to produce a computer-implemented process such that the instructions which execute via the one or more processors of the computer or other programmable data processing devices create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0148] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0149] The flow diagrams and the block diagrams in the drawings are presented to illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to the present application. In this regard, each block in the flow diagrams and the block diagrams can represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical functions ("instructions"). In some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks can sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and / or flow diagrams, and combinations thereof, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and

[0150] Example embodiments have been disclosed herein and, although specific terms are employed, they are used in a generic and descriptive sense only and not for purposes of limitation. In some instances, it will be apparent to those skilled in the art that features, characteristics or aspects described in relation to one embodiment can be applied to other embodiments, unless otherwise clearly stated. It will also be apparent to those skilled in the art that various modifications can be made to the described embodiments without departing from the scope of the invention as defined by the appended claims.

Claims

1. A parking area detection method based on cascade forest and stacked ensemble learning, characterized in that: The following steps are involved: S1. Obtaining a feature matrix of a road image; S2. Inputting the feature matrix into a cascade forest model, wherein each layer of the cascade forest model includes a fixed number of multiple tree-structured base learners, and the cascade forest model outputs multiple vehicle drivable area segmentation results corresponding to the multiple tree-structured base learners; S3. Use the stacked ensemble learning algorithm to learn the output of the cascade forest model to obtain the final vehicle drivable area segmentation result.

2. The method according to claim 1, characterized in that Step S1 includes: performing feature extraction on the road image, wherein the extracted features include at least one of color features, texture features, edge features, and contour features; Build based on feature extraction results dimensional feature matrix, where Represents the number of pixel samples in the road image, The dimension of image features representing pixel samples.

3. The method according to claim 2, characterized in that The extraction of the color feature comprises the following steps: I. Color histogram feature extraction; Count the red, green and blue channel values ​​of each pixel in the road image and distribute them in a histogram. By calculating the frequency of occurrence of pixel values ​​in each channel, we can obtain information about the color distribution of the road image. II. Color space feature extraction; Convert the road image from RGB to hue, saturation, and value color space and extract HSV information. The calculation formula is as follows: Where, 、 、 Represents the red, green, and blue channel values ​​of the pixel respectively; Indicates hue; Indicates saturation; Indicates brightness.

4. The method according to claim 2, characterized in that The extraction of the texture features comprises the following steps: I. Gray Level Co-occurrence Matrix (GLCM) feature extraction; By counting the co-occurrence frequency of grayscale values ​​between pixel pairs in the image, the texture information of the vehicle's drivable area is extracted. The calculation formula is as follows: Where, is the gray level in the image; According to the angle and distance Determined horizontal displacement; According to the angle and distance Determined vertical displacement; is the size of the image; II. Local binary pattern (LBP) feature extraction; By comparing the grayscale values ​​of pixels with those of corresponding neighboring pixels, the texture information of the vehicle drivable area in the neighborhood of the pixel point is extracted. The calculation formula is as follows: in, are the coordinates of the center pixel; The coordinates are The number of neighboring pixels of the central pixel, It is The grayscale value of pixels in the area, and is an integer; is the gray value of the center pixel; function Defined as: 。 5. The method according to claim 2, characterized in that The extraction of the edge features comprises the following steps: identifying the discontinuity of pixel values ​​by using the Canny edge detection algorithm to determine the edge position; The calculation formula is as follows: Where, is the magnitude of the gradient; is the horizontal component of the gradient; is the longitudinal component of the gradient; Characterize the direction of the gradient; The local maximum gradient value is retained for the calculated gradient, and the edge is determined using the double threshold method. Finally, the edge pixels are connected to obtain the edge features of the vehicle's drivable area.

6. The method according to claim 2, characterized in that The extraction of the outline features comprises the following steps: Describing the contour features in the image by shape descriptors, thereby extracting the geometric features of the object to represent the shape of the object; the shape descriptors include Hu moments and Zernike moments; The Hu moment is calculated as follows: Where, is the original moment of the image, is the order; is the coordinate in the image The pixel value at ; is the central moment of the image; are the coordinates of the image's centroid, , ; The calculation formula of Zernike moment is as follows: Where, yes Second-rate Zernike moments of order, is the radial order, is the angular order, satisfying and is an even number; are Zernike orthogonal polynomials.

7. The method according to claim 2, characterized in that In step S3, a support vector machine is used as a meta-learner to learn the output of the cascade forest model.

8. The method according to any one of claims 1 to 7, characterized in that The multiple tree structure-based learners include: 2 random forest learners and 2 XGBoost learners.

9. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 8.

10. A computer-readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images