Video-based parking space occupancy detection system and method, and artificial intelligence training method therefor

Abstract

Provided is an artificial intelligence training method for video-based parking space occupancy detection. The method comprises the steps of: detecting objects from a video in which a plurality of parking spaces in a parking lot are captured; generating training data based on cumulative feature information which is pieces of feature information, corresponding to the objects detected from the respective images constituting the video, accumulated for a predetermined time; and training, on the basis of the training data, artificial intelligence for detecting whether the objects occupy the parking spaces.

Description

Video-based parking space occupancy detection system and method, and artificial intelligence learning method thereof

[0001] The present invention relates to an image-based parking space occupancy detection system and method, and an artificial intelligence learning method thereof.

[0002] Traditional methods for finding parking spaces in parking lots have largely been manual. This means that to find a space, you have to drive around the lot, visually searching for it.

[0003] However, recent advancements in sensor technology, image processing, and artificial intelligence have led to the development and application of technologies that automatically detect parking spaces. For example, technologies using ultrasonic sensors to detect parking spaces and video analysis to detect parking spaces are being applied.

[0004] However, sensor-based parking detection technology requires installation in each parking space, resulting in significant installation and maintenance costs. Furthermore, even with existing video analysis technology, recognizing a vehicle's presence in a parking space requires prior manual delineation of each parking space and the subsequent designation of the appropriate parking spaces.

[0005] The problem to be solved by the present invention is to provide an image-based parking space occupancy detection method and system that can detect whether a parking space is occupied through artificial intelligence based on the characteristic information of objects accumulated over a certain period of time without zoning the parking space in the parking lot, and also to provide an artificial intelligence learning method therefor.

[0006] However, the problems to be solved by the present invention are not limited to the problems described above, and other problems may exist.

[0007] According to a first aspect of the present invention for solving the above-described problem, an artificial intelligence learning method for detecting parking space occupancy based on an image includes the steps of: detecting an object from an image in which a plurality of parking spaces in a parking lot are captured; generating learning data based on accumulated feature information obtained by accumulating feature information corresponding to an object detected from each image constituting the image for a predetermined period of time; and learning an artificial intelligence for detecting whether the object occupies the parking space based on the learning data.

[0008] In some embodiments of the present invention, the step of generating the learning data may include the step of setting a predetermined location from the object area of ​​the detected object as feature information; and the step of digitizing the feature information of each image to express it as feature information, setting an arbitrary axis for each feature information, and then generating statistical distribution information that is accumulated and densely displayed for a predetermined period of time on a relative coordinate system formed by the axis as the accumulated feature information.

[0009] In some embodiments of the present invention, the step of generating statistical distribution information in which feature information of each image is accumulated over a certain period of time and densely displayed as the accumulated feature information may generate statistical distribution information of n variables (n is a natural number) using at least one of coordinate information at a specific location of the object area, width information of the object area, height information, object type information, and object shape information as the accumulated feature information.

[0010] In some embodiments of the present invention, the step of learning the artificial intelligence may include a step of inputting the feature information included in the accumulated feature information into the artificial intelligence and clustering it according to an arbitrary criterion to generate candidate cluster information; and a step of re-clustering the candidate cluster information through the artificial intelligence to generate cluster information corresponding to the parking space.

[0011] In some embodiments of the present invention, the step of generating the candidate cluster information may generate the feature information corresponding to the predetermined distribution as candidate cluster information when the feature information included in the accumulated feature information is densely packed within a predetermined distribution by a preset number or more.

[0012] In some embodiments of the present invention, the step of generating cluster information corresponding to the parking space may include the step of matching at least one candidate cluster information with the location of an actual parking space prepared in advance; and the step of generating at least one candidate cluster information matched with the location of the same parking space as cluster information corresponding to the parking space.

[0013] In addition, a method for detecting parking space occupancy based on an image according to a second aspect of the present invention includes the steps of: detecting an object included in an image in an image of a plurality of parking spaces in a parking lot; extracting feature information corresponding to the detected object; and inputting the feature information into an artificial intelligence trained in advance to detect whether the parking space is occupied, thereby detecting whether the parking space is occupied by an object. In this case, the artificial intelligence may be trained using learning data based on accumulated feature information obtained by accumulating feature information corresponding to objects detected from each image constituting an image of the parking space over a certain period of time.

[0014] In addition, a video-based parking space occupancy detection system according to a third aspect of the present invention includes a communication module that receives a video of a plurality of parking spaces in a parking lot from a camera, a memory in which a program for generating learning data based on the video and learning artificial intelligence for detecting whether the parking space is occupied based on the learning data is stored, and a processor that detects an object from the video by executing the program stored in the memory, generates the learning data based on accumulated feature information corresponding to the object detected from each image constituting the video for a predetermined period of time, and performs unsupervised learning of the artificial intelligence based on the learning data.

[0015] In addition, a computer program according to another aspect of the present invention is coupled with a computer as hardware to execute the image-based parking space occupancy detection artificial intelligence learning method and detection method, and is stored in a computer-readable recording medium.

[0016] Other specific details of the present invention are included in the detailed description and drawings.

[0017] According to the present invention described above, it is possible to detect whether a parking space is occupied by a large number of vehicles at once, that is, whether a vehicle is parked, using only camera images capturing a plurality of parking spaces.

[0018] In this process, unlike conventional image processing technology, there is no need to input parking space partition information into the system in advance to have the system recognize the location information of the parking space, and there is an advantage in that it can recognize parking spaces based on probability and detect whether a vehicle is parked through the unsupervised learning process of artificial intelligence.

[0019] In addition, by configuring various variables as learning data, it has the advantage of enabling more accurate parking space detection and detection of parking availability by considering the type and shape of objects.

[0020] The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

[0021] FIG. 1 is a flowchart of an artificial intelligence learning method for image-based parking space occupancy detection according to one embodiment of the present invention.

[0022] FIG. 2a and FIG. 2b are drawings for explaining object detection results and feature information in one embodiment of the present invention.

[0023] Figures 3a to 3c are drawings for explaining a process of generating learning data in one embodiment of the present invention.

[0024] FIG. 4 is a diagram for explaining candidate cluster information in one embodiment of the present invention.

[0025] Figure 5 is a diagram for explaining cluster information in one embodiment of the present invention.

[0026] FIG. 6 is a diagram for explaining a process of determining whether an object occupies a parking space by applying learned artificial intelligence in one embodiment of the present invention.

[0027] FIG. 7 is a block diagram of an image-based parking space occupancy detection system according to one embodiment of the present invention.

[0028] The advantages and features of the present invention, and the methods for achieving them, will become clearer with reference to the embodiments described in detail below together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are provided solely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the present invention, and the present invention is defined solely by the scope of the claims.

[0029] The terminology used herein is for the purpose of describing embodiments only and is not intended to limit the present invention. In this specification, the singular also includes the plural unless specifically stated otherwise. As used herein, the terms "comprises" and / or "comprising" do not exclude the presence or addition of one or more other components in addition to the mentioned components. Like reference numerals refer to like components throughout the specification, and "and / or" includes each and any combination of one or more of the mentioned components. Although "first", "second", etc. are used to describe various components, these components are not limited by these terms. These terms are only used to distinguish one component from another. Therefore, it should be understood that a first component mentioned below may also be a second component within the technical spirit of the present invention.

[0030] Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used in their common sense to those skilled in the art to which the present invention pertains. Furthermore, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise.

[0031] Hereinafter, an image-based parking space occupancy detection artificial intelligence learning method and detection method according to one embodiment of the present invention will be described with reference to FIGS. 1 to 6.

[0032] Figure 1 is a flowchart of an artificial intelligence learning method for image-based parking space occupancy detection according to one embodiment of the present invention. While each step illustrated in Figure 1 may be understood to be performed by the image-based parking space occupancy detection system (100) described below, the present invention is not necessarily limited thereto.

[0033] First, objects are detected from images of multiple parking spaces within a parking lot (S110). In one embodiment of the present invention, the target parking space can be an image of a parking lot with partitioned parking spaces, but images of parking lots without partitioned parking spaces are also applicable. This is possible through a process of clustering feature information in the artificial intelligence learning stage described below.

[0034] In one embodiment, an object detection algorithm, one of the image processing techniques, can be used to detect vehicles within a parking lot. The object detection algorithm may be a deep learning-based algorithm. One embodiment of the present invention can automatically detect objects, such as vehicles, in an image using the object detection algorithm. Furthermore, a process of identifying the location and size of the object can be performed to detect the object in the image.

[0035] Next, learning data is generated based on accumulated feature information corresponding to objects detected from each image constituting the video over a certain period of time (S120).

[0036] FIG. 2a and FIG. 2b are drawings for explaining object detection results and feature information in one embodiment of the present invention.

[0037] When an object is detected through an object detection algorithm, the object can be displayed together with object area information (P1) as shown in Fig. 2a. Location information of the object detected in this manner can be extracted. In this case, one embodiment of the present invention can extract a specific point (x, y) of the object area as feature information as the location information as shown in Fig. 2b. However, the feature information is not necessarily limited thereto, and the feature information can include up, down, left, and right length and height values, center point coordinates, corner point coordinates, color, and type information. This feature information may extract only a specific coordinate value. Alternatively, the feature information may extract a plurality of coordinate values, and may be extracted by combining a plurality of feature information having different locations and characteristic information.

[0038] In addition, one embodiment of the present invention can extract different feature information for each location of a parking space constituting a parking lot. This is because the location, shape, etc. of the detected object area are different depending on the location of the parking space on the image, particularly the direction. Therefore, as an example, based on the camera image in FIG. 2A, for the right parking spaces, the coordinates of the lower left point of the object area information can be extracted as feature information, for the front parking spaces, the coordinates of the lower center point can be extracted as feature information, and for the left parking spaces, the coordinates of the lower right point can be extracted as feature information.

[0039] The extracted feature information can be used as representative information for the detected object. After training, the feature information can be utilized for parking management, such as detecting vehicle occupancy in parking spaces and determining vehicle parking status. As mentioned earlier, a parking space can be a conventional space where parking is performed without a designated parking area.

[0040] Figures 3a to 3c are drawings for explaining a process of generating learning data in one embodiment of the present invention.

[0041] After feature information is extracted from the object area of ​​the detected object, the feature information of each image can be accumulated by setting an arbitrary axis for the feature information, and then generating statistical distribution information (e.g., histogram information) that is accumulated and displayed over a certain period of time on a relative coordinate system formed by the axis.

[0042] At this time, one embodiment of the present invention can use not only coordinate information at a specific location of an object area as accumulated feature information, but also statistical distribution information of n variables (n is a natural number) that uses at least one of width information, height information, object type information, and object information of an object area as a variable.

[0043] For example, two-variable statistical distribution information that applies coordinate information (x, y) of a specific reference point can be configured as learning data. As another example, one-variable statistical distribution information that applies only information collected based on one of the axes (e.g., x-axis) can be configured as learning data. In addition, n-variable statistical distribution information can be configured by combining the height (h), width (w), or diagonal length (d) of the object area.

[0044] Furthermore, one embodiment of the present invention may further combine object type information and object shape information to form statistical distribution information of n variables. Considering object type information and object shape information as variables is intended to more accurately recognize and detect parked vehicles, as various types and shapes of vehicles exist within parking lot images.

[0045] Specifically, one embodiment of the present invention, when considering object type information as a variable, can more accurately distinguish various vehicle types within a parking lot. For example, different variable values ​​can be applied to vehicle types such as passenger cars, SUVs, and trucks. Since each vehicle type has different characteristics, such as parking location and size, considering object type as a variable has the advantage of enabling more accurate parking space detection and object occupancy detection.

[0046] In addition, in one embodiment of the present invention, when considering the shape information of an object as a variable, more accurate parking space detection and detection of object occupancy can be enabled by applying different variables according to the shape of a parked vehicle. That is, since the direction and location of parking differ depending on the shape of the parked vehicle, there is an advantage in that more accurate parking space detection and detection of object occupancy are enabled by applying different variables to long passenger cars (such as mid- to large-sized sedans) and short passenger cars (such as compact cars).

[0047] In this way, one embodiment of the present invention further considers object type and shape information, thereby further distinguishing between various types and shapes of vehicles within a parking lot, thereby enabling more accurate detection of whether various vehicles occupy parking spaces within the parking space. While the candidate cluster information may be generated based on the object type and shape, ultimately only one cluster information is generated for each parking space, and by generating different candidate cluster information, a variety of information can be provided.

[0048] Meanwhile, one embodiment of the present invention may perform a data preprocessing process before generating statistical distribution information for feature information collected over a certain period of time. In this case, one embodiment of the present invention may generate statistical distribution information after performing preprocessing processes such as noise removal and pruning on the feature information.

[0049] In addition, one embodiment of the present invention can diagram the distribution of feature information when generating statistical distribution information.

[0050] In one embodiment, statistical distribution information can further provide the density of feature information distribution, and this density can be provided in a graphical form. For example, the density can be displayed with different colors for each level. In another example, the density can be displayed in the form of a contour line, as shown in Figure 3c. In this way, statistical distribution information can provide density information, which can be generated as cluster information during the AI ​​learning process.

[0051] Next, artificial intelligence is trained to detect parking occupancy for an object's parking space based on learning data (S130).

[0052] FIG. 4 is a diagram illustrating candidate cluster information in one embodiment of the present invention. FIG. 5 is a diagram illustrating cluster information in one embodiment of the present invention.

[0053] One embodiment of the present invention can perform unsupervised learning of artificial intelligence through accumulated feature information based on statistical distribution information in the previous step.

[0054] At this time, the present invention allows the artificial intelligence to perform unsupervised learning, i.e., train itself. First, feature information included in the accumulated feature information is input into the artificial intelligence, and the artificial intelligence can generate candidate cluster information that is clustered by itself according to arbitrary criteria. In Figure 4, candidate cluster information is indicated by A to W.

[0055] Here, the arbitrary criterion may be a criterion discovered or set by the artificial intelligence itself. In one embodiment, if the feature information included in the accumulated feature information is densely clustered within a certain distribution by a preset number or more, or if the population distribution estimated through the distribution meets conditions such as a certain standard deviation, Z, or confidence interval, the artificial intelligence may generate feature information corresponding to the certain distribution as candidate cluster information. In another embodiment, the artificial intelligence may calculate the similarity between the feature information included in the accumulated feature information, and cluster feature information having a similarity higher than a preset threshold to generate candidate cluster information. That is, in one embodiment of the present invention, the artificial intelligence may generate candidate cluster information by performing automatic pattern recognition and statistical classification on data through unsupervised learning. At this time, it goes without saying that the criterion used may be set in various ways depending on the characteristics of the algorithm and the characteristics of the data.

[0056] After candidate cluster information is generated, one embodiment of the present invention can re-cluster the candidate cluster information using artificial intelligence to generate cluster information corresponding to parking spaces. The example in Fig. 5 illustrates the generation of a total of 10 cluster information. At this time, candidate cluster information A to D were generated as cluster information 1 (1-1 to 1-4), and candidate cluster information E was generated as cluster information 2 (2-1).

[0057] In the artificial intelligence learning process for generating cluster information using candidate cluster information, the location of an actual parking space prepared in advance is first matched with at least one candidate cluster information, and at least one candidate cluster information matched or tagged with the location of the same parking space can be generated as cluster information corresponding to the parking space.

[0058] The actual locations of parking spaces can be obtained from images or existing parking lot maps, and these can be matched with candidate cluster information. This matching process can be achieved by detecting common features or similar patterns between actual parking spaces and candidate cluster information.

[0059] If the parking lot is not a typical, partitioned parking lot, but rather an un-partitioned one, actual parking space information may be pre-defined or collected and analyzed based on existing vehicle parking patterns. Alternatively, AI may match the information through comparison with surrounding environmental features (e.g., buildings, trees, roads, and other structures).

[0060] Meanwhile, in cases where there are no parking space segments, it is desirable to generate various types of cluster information by applying information such as multivariate histograms and scatter plots when applying the aforementioned statistical distribution information. For example, by setting object type and shape information as variables in the statistical distribution information, the AI ​​can learn to match candidate cluster information with parking spaces that accommodate vehicles of various sizes and types.

[0061] In addition, if the number of candidate cluster information is not large, the process of matching candidate cluster information with actual parking spaces may be performed manually by the administrator.

[0062] FIG. 6 is a diagram for explaining a process of determining whether an object occupies a parking space by applying learned artificial intelligence in one embodiment of the present invention.

[0063] After the learning of the artificial intelligence is completed, objects included in images in a video of multiple parking spaces in a parking lot are detected in real time (S140), and feature information of the detected objects is extracted (S150).

[0064] Next, by inputting feature information into the learned artificial intelligence, the result of detecting whether an object is occupied in a parking space can be output (S160).

[0065] Meanwhile, when training AI, there are frequent instances where vehicles temporarily park in front of entrances to underground parking lots, even though they are not designated parking spaces. In such cases, if this feature information is directly integrated into cumulative feature information, clusters of non-parking spaces may be generated as parking spaces, potentially leading to errors.

[0066] To address this issue, one embodiment of the present invention may further configure vehicle parking time information as learning data when configuring variables constituting statistical distribution information. Then, the average parking time of vehicles corresponding to characteristic information within the cluster is calculated based on the generated cluster information, and if the average parking time satisfies or exceeds a preset minimum parking time, the cluster information may be determined to be a parking space. At this time, the minimum parking time information may be set to a lower specific % or a minimum value of the parking time statistics of vehicles determined to be parking spaces, and the time corresponds to the time until the same vehicle exits the parking space again after parking is completed.

[0067] Meanwhile, in the above description, steps S110 to S160 may be further divided into additional steps or combined into fewer steps, depending on the implementation of the present invention. Furthermore, some steps may be omitted as needed, and the order of the steps may be changed. Furthermore, even if other omitted details are present, the contents of FIGS. 1 to 6 may also be applied to the video-based parking space occupancy detection system (100) of FIG. 7, which will be described later.

[0068] FIG. 7 is a block diagram of an image-based parking space occupancy detection system (100) according to one embodiment of the present invention.

[0069] An image-based parking space occupancy detection system (100) according to one embodiment of the present invention includes a communication module (110), a memory (120), and a processor (130).

[0070] The communication module (110) receives images of multiple parking spaces in a parking lot from a camera. The communication module (110) may include both a wired communication module and a wireless communication module. The wired communication module may be implemented as a power line communication device, a telephone line communication device, a cable home (MoCA), Ethernet, IEEE1294, an integrated wired home network, and an RS-485 control device. In addition, the wireless communication module may be implemented as a wireless LAN (WLAN), Bluetooth, HDR WPAN, UWB, ZigBee, Impulse Radio, 60GHz WPAN, Binary-CDMA, wireless USB technology, and wireless HDMI technology.

[0071] The memory (120) stores a program for generating learning data based on images and learning artificial intelligence to detect whether a parking space is occupied based on the learning data, and the processor (130) executes the program stored in the memory (120).

[0072] Here, memory (120) refers to a general term for non-volatile storage devices and volatile storage devices that maintain stored information even when power is not supplied. For example, memory (120) may include NAND flash memory such as a compact flash (CF) card, an SD (secure digital) card, a memory stick, a solid-state drive (SSD), and a micro SD card, magnetic computer storage devices such as a hard disk drive (HDD), and optical disc drives such as a CD-ROM and a DVD-ROM.

[0073] The processor (130) detects an object from an image, generates learning data based on accumulated feature information corresponding to the object detected from each image constituting the image over a certain period of time, and performs unsupervised learning of artificial intelligence based on the learning data.

[0074] The image-based parking space occupancy detection artificial intelligence learning method and detection method according to one embodiment of the present invention described above can be implemented as a program (or application) to be executed in combination with a computer as hardware and stored in a medium.

[0075] The above-described program may include codes coded in a computer language, such as C, C++, JAVA, Ruby, Swift, JavaScript, Python, or machine language, that can be read by the processor (CPU) of the computer through the device interface of the computer, so that the computer reads the program and executes the methods implemented as a program. Such codes may include functional codes related to functions that define functions necessary to execute the methods, and may include control codes related to execution procedures necessary for the processor of the computer to execute the functions according to a predetermined procedure. In addition, such codes may further include memory reference-related codes regarding which location (address address) of the internal or external memory of the computer should reference additional information or media necessary for the processor of the computer to execute the functions. In addition, if the processor of the computer needs to communicate with any other computer or server located remotely in order to execute the functions, the code may further include communication-related code regarding how to communicate with any other computer or server located remotely using the communication module of the computer, and what information or media to send and receive during communication.

[0076] The above storage medium refers to a medium that stores data semi-permanently and can be read by a device, rather than a medium that stores data for a short period of time, such as a register, cache, or memory. Specifically, examples of the storage medium include, but are not limited to, ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical data storage device. That is, the program can be stored in various recording media on various servers that the computer can access or in various recording media on the user's computer. In addition, the medium can be distributed across network-connected computer systems, so that computer-readable code can be stored in a distributed manner.

[0077] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will readily appreciate that the present invention can be readily modified into other specific forms without altering the technical spirit or essential characteristics of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single entity may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined manner.

[0078] The scope of the present invention is indicated by the claims described below rather than the detailed description above, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention.

Claims

1. In a method performed by a computer, A step of detecting an object from an image of multiple parking spaces in a parking lot; A step of generating learning data based on accumulated feature information corresponding to objects detected from each image constituting the above image over a certain period of time; and A step of learning artificial intelligence to detect whether the object occupies the parking space based on the learning data, An artificial intelligence learning method for video-based parking space occupancy detection.

2. In paragraph 1, The steps for generating the above learning data are: A step of setting a predetermined position coordinate and characteristic from the object area of ​​the detected object as feature information; and A step of generating statistical distribution information that is densely displayed by accumulating for a certain period of time on a relative coordinate system formed by the axis after setting an arbitrary axis for the characteristic information of each of the images above, as the accumulated characteristic information. An artificial intelligence learning method for video-based parking space occupancy detection.

3. In paragraph 2, The step of generating statistical distribution information in which feature information of each of the above images is accumulated over a certain period of time and densely displayed as the accumulated feature information is as follows. The statistical distribution information of n variables (n is a natural number) using at least one of coordinate information at a specific location point of the object area, width information of the object area, height information, object type information, and object shape information as a variable is generated as the accumulated feature information. An artificial intelligence learning method for video-based parking space occupancy detection.

4. In paragraph 1, The steps of learning the above artificial intelligence are: A step of inputting the feature information included in the above-mentioned accumulated feature information into the artificial intelligence and generating candidate cluster information clustered according to arbitrary criteria; and A step of re-clustering the above candidate cluster information using the artificial intelligence to generate cluster information corresponding to the parking space, An artificial intelligence learning method for video-based parking space occupancy detection.

5. In paragraph 4, The step of generating the above candidate cluster information is: When the feature information included in the above-mentioned accumulated feature information is densely packed in a predetermined number or is greater than a predetermined statistical value within a certain distribution, feature information corresponding to the predetermined distribution is generated as candidate cluster information. An artificial intelligence learning method for video-based parking space occupancy detection.

6. In paragraph 4, The step of generating cluster information corresponding to the above parking space is: A step of matching the location of a pre-prepared actual parking space with at least one candidate cluster information; and A step of generating at least one candidate cluster information matching the location of the same parking space as cluster information corresponding to the parking space, An artificial intelligence learning method for video-based parking space occupancy detection.

7. In a method performed by a computer, A step of detecting an object included in an image in a video of multiple parking spaces in a parking lot; A step of extracting feature information corresponding to the detected object; and A step of inputting the above characteristic information into an artificial intelligence that has been trained in advance to detect whether the parking space is occupied by an object, and detecting whether the parking space is occupied by an object, The above artificial intelligence is learned using learning data based on accumulated feature information corresponding to objects detected from each image constituting the video of the above parking space over a certain period of time. A video-based parking space occupancy detection method.

8. A communication module that receives images of multiple parking spaces in a parking lot from a camera. A memory storing a program for generating learning data based on the above image and learning artificial intelligence to detect whether the parking space is occupied based on the learning data, and By executing the program stored in the memory, the processor detects an object from the image, generates the learning data based on the accumulated feature information corresponding to the object detected from each image constituting the image for a certain period of time, and performs unsupervised learning of the artificial intelligence based on the learning data. Video-based parking space occupancy detection system.

Images