An intelligent management system for regional properties

Abstract

The application relates to an area property intelligent management system, which comprises multiple image sensors, multiple guides and a control system. The multiple image sensors are arranged in a closed area, and the image sensors are used for acquiring image information in a coverage range. The multiple guides are arranged in the closed area, and the guides are used for generating guide information according to a vehicle object. The control system is in data communication with the image sensors and the guides, the control system is used for generating a guide path matched with the vehicle object, tracking the vehicle object according to image information fed back by the image sensors and sending guide information to the vehicle object by using the guides. The area property intelligent management system disclosed by the application uses the mode of image recognition, guide and destination path planning to guide the user to quickly complete vehicle parking, and the destination is referred to during path planning, so that the user vehicle can be as close to the destination as possible.

Description

Technical Field

[0001] This application relates to the field of smart property management technology, and in particular to a regional property intelligent management system. Background Technology

[0002] With the rapid development of technology, the property management industry is undergoing profound changes. As a key tool for improving property management efficiency and optimizing service quality, digital property management systems have demonstrated powerful management capabilities in various scenarios, such as security management, infrastructure management, humanized management, and resident service management.

[0003] In actual commercial use, it has been found that the demand for vehicle management is increasing, mainly focusing on vehicles, parking spaces, and visitor vehicles. This is because commercial parking lots are mostly underground, with a large number of parking spaces and high traffic volume. In addition, underground parking lots have practical problems such as communication signals and positioning, and untimely signal updates, which make it impossible to effectively guide vehicles, resulting in a poor parking experience for customers. There is no connection between the parking location and the elevator to enter the mall, and customers still need to find their own way. Summary of the Invention

[0004] This application provides a regional property intelligent management system that uses image recognition of vehicles, guide devices, and destination route planning to guide users to quickly park their vehicles. At the same time, the system takes the destination into account during route planning so that the user's vehicle can get as close to the destination as possible.

[0005] The above-mentioned objective of this application is achieved through the following technical solution:

[0006] This application provides a regional property intelligent management system, including:

[0007] Multiple image sensors are set up within an enclosed area; the image sensors are used to acquire image information within the coverage area.

[0008] Multiple guides are set up within an enclosed area. The guides are used to generate guidance information based on the vehicle object.

[0009] The control system communicates with the image sensor and the guide. The control system generates a guide path that matches the vehicle object, tracks the vehicle object based on the image information fed back by the image sensor, and sends guide information to the vehicle object using the guide.

[0010] In one possible implementation of this application, the control system tracks the vehicle object based on image information fed back by the image sensor, including:

[0011] The image sensor is invoked along the guide path, and its orientation is adjusted according to the guide path.

[0012] Obtain image information fed back by the invoked image sensor;

[0013] Identify vehicle objects included in the image information;

[0014] Determine the position of the vehicle object on the guide path based on the position of the vehicle object, and record it as the current position of the vehicle object;

[0015] Select a guide based on the current location of the vehicle object and send guidance information to the selected guide;

[0016] The vehicle objects included in the image information are identified using license plate recognition and comparison recognition methods.

[0017] When the lighting conditions meet the requirements, license plate recognition is used for identification; when the lighting conditions do not meet the requirements, comparison recognition is used for identification.

[0018] In one possible implementation of this application, identifying vehicle objects included in image information using a contrast recognition method includes establishing a feature set, which includes:

[0019] Acquire the first reference image of the vehicle object when it reaches the entry position;

[0020] Establish a first reference line on the first reference image and establish feature points on the first reference line;

[0021] A feature set belonging to the first reference line is created based on the first reference line. The feature set includes the starting point position of the first reference line, the ending point position of the first reference line, the length of the first reference line, the angle of the first reference line, the position of the feature points, and the distance between the feature points.

[0022] In one possible implementation of this application, identification using comparison recognition includes:

[0023] Obtain a second reference image of the vehicle object along the guide path that matches the vehicle object;

[0024] Local feature information of the vehicle object is obtained from the second reference image;

[0025] The relative position of the image sensor and the vehicle object corresponding to the second reference image is determined using local feature information.

[0026] Select a suitable image sensor based on the relative position and acquire a second reference image again;

[0027] Extract the first reference line and its feature set from the second reference image, and denote it as the contrast feature set.

[0028] Compare and contrast the feature set and the feature set to obtain the comparison result;

[0029] The vehicle is determined based on the comparison results.

[0030] In one possible implementation of this application, extracting the first reference line and its feature set from the second reference image includes:

[0031] Features belonging to the vehicle object are obtained on the first reference image. The features belonging to the vehicle object include the starting point position of the first reference line and / or the ending point position of the first reference line.

[0032] Extract the feature contours belonging to the vehicle object and move the feature contours to the corresponding positions on the second reference image;

[0033] Adjust the state of the feature contour on the second reference image so that the overlap between the feature contour and the contour at the corresponding position on the second reference image meets the requirements;

[0034] A second reference line is established using the contour at the corresponding position on the second reference image;

[0035] Use the second reference line to generate a set of contrast features for the second reference line.

[0036] In one possible implementation of this application, adjusting the state of the feature contour on the second reference image includes zooming in, zooming out, and rotating.

[0037] In one possible implementation of this application, generating the contrast feature set of the second reference line using the second reference line includes:

[0038] Use the pixels on the second reference line to generate a pixel difference curve or a quadratic pixel difference curve;

[0039] Obtain the curvature change points of the pixel difference curve or the quadratic curve of pixel difference, and record the curvature change points as the reference positions of feature points and the reference distances between feature points;

[0040] The matching result is obtained by matching the feature point reference position with the feature point position on the first reference line based on the distance between the feature point reference position and the starting point and / or the ending point of the second reference line;

[0041] Generate the corresponding reference distance between feature points based on the matching results.

[0042] In one possible implementation of this application, there are multiple first reference lines.

[0043] In one possible implementation of this application, at least two first reference lines are parallel and at least two first reference lines intersect.

[0044] In one possible implementation of this application, there are at least two parallel first reference lines. The starting points of the two first reference lines are located on the same sub-object of the vehicle object, and the ending points of the two first reference lines are located on the same sub-object of the vehicle object. Attached Figure Description

[0045] Figure 1 This is a schematic block diagram of a regional property intelligent management system provided in this application.

[0046] Figure 2 This is a schematic diagram of a guiding path provided in this application.

[0047] Figure 3 This is a schematic diagram of a bootloader displaying boot information provided in this application.

[0048] Figure 4 This is a schematic diagram of another type of bootloader displaying boot information provided in this application.

[0049] Figure 5 This is a schematic diagram of obtaining a first reference image provided in this application.

[0050] Figure 6 This is a schematic diagram of establishing a first reference line on a first reference image and establishing feature points on the first reference line, as provided in this application.

[0051] Figure 7 This is a schematic diagram of a first reference image provided in this application having multiple first reference lines. Detailed Implementation

[0052] The technical solutions in this application will be further described in detail below with reference to the accompanying drawings.

[0053] This application discloses a regional property intelligent management system; please refer to [link / reference]. Figure 1 In some examples, the regional property intelligent management system disclosed in this application includes an image sensor 1, a guide 2, and a control system 3. The image sensor 1 and the guide 2 are both deployed inside a closed area, which refers to an underground parking lot. Generally, the image sensor 1 and the guide 2 are deployed on the top of the underground parking lot. The image sensor 1 and the guide 2 generally use wired data communication or local wireless communication network to communicate with the control system 3.

[0054] Image sensor 1 and guide 2 are typically deployed at the entrance, at the intersection of the exit and the road.

[0055] Image sensor 1 is used to acquire image information within its coverage area, guide 2 is used to generate guidance information based on the vehicle object, and control system 3 is used to generate a guidance path that matches the vehicle object, such as... Figure 2 As shown, Figure 2 There are four elevators in the building. Generally speaking, the vehicle should be parked near the elevator closest to the destination, or it can be a staircase.

[0056] Based on the image information fed back by image sensor 1, the vehicle object is tracked, and guide information is sent to the vehicle object using guide 2, such as... Figure 3 As shown, the guidance information generally consists of two parts: the license plate number and an arrow. When there are multiple pieces of guidance information, the guide 2 can use a multi-row, multi-column display method, such as... Figure 4 As shown.

[0057] The control system 3 tracks the vehicle object based on the image information fed back by the image sensor 1. The specific process is as follows:

[0058] S101, Invoke image sensor 1 on the guide path and adjust the orientation of image sensor 1 according to the guide path;

[0059] S102, Obtain image information fed back by the called image sensor 1;

[0060] S103, Identify the vehicle objects included in the image information;

[0061] S104, Determine the position of the vehicle object on the guide path based on the position of the vehicle object, and record it as the current position of the vehicle object;

[0062] S105, Select guide 2 according to the current location of the vehicle object and send guidance information to the selected guide 2;

[0063] The vehicle objects included in the image information are identified using license plate recognition and comparison recognition methods.

[0064] When the lighting conditions meet the requirements, license plate recognition is used for identification; when the lighting conditions do not meet the requirements, comparison recognition is used for identification.

[0065] When a vehicle enters the underground parking lot, it first selects its destination at the entrance. The destination can be selected in various ways, such as by operating a smart terminal, being guided by staff, or selecting it on a service terminal. There are no restrictions on this method.

[0066] Next, a guide path is generated based on the destination. After the guide path is generated, step S101 is executed. In this step, image sensor 1 is invoked on the guide path and the orientation of image sensor 1 is adjusted according to the guide path. The purpose of adjusting the orientation of image sensor 1 is to obtain an image including vehicle objects as quickly as possible.

[0067] Of course, it is also necessary to pay attention to the actual position of the vehicle object here. When calling image sensor 1 on the guide path, it is generally appropriate to call it according to the actual position of the vehicle object, rather than calling it all directly.

[0068] In some possible implementations, one or two image sensors are typically invoked in the forward direction of the vehicle object.

[0069] In step S102, image information fed back by the invoked image sensor 1 is acquired, and then in S103, vehicle objects included in the image information are identified.

[0070] Next, step S104 is executed. In step S104, the position of the vehicle object on the guide path is determined based on the position of the vehicle object and recorded as the current position of the vehicle object. The current position of the vehicle object can be determined by referring to the position of image sensor 1, because the position and orientation of image sensor 1 are fixed.

[0071] When image sensor 1 uses a spherical image sensor, it is necessary to use markers in the surrounding environment (usually located on a wall) to determine the orientation.

[0072] In step S105, a guide 2 is selected based on the current location of the vehicle object and guidance information is sent to the selected guide 2. In other words, guidance information is sent to the guide 2 in a moderate manner, rather than sending guidance information directly to all the guides 2 involved.

[0073] Generally, guidance information is sent to the guide 2 within a 5-10 meter range along the direction of the vehicle's movement.

[0074] There are two methods used to identify vehicles in image information: license plate recognition and comparison recognition. The selection criterion for these two methods is whether the lighting environment meets the requirements. It should be understood that when the lighting environment is poor, the license plate information in the obtained image may be incomplete and difficult to recognize.

[0075] As for the selection of the lighting environment, it is generally fixed when deploying image sensor 1 and guide 2. Of course, it can also be determined by deploying a certain number of light intensity sensors, and then the license plate recognition method or the comparison recognition method can be selected based on the feedback data of the light intensity sensors.

[0076] After all, considering the amount and difficulty of data processing, the implementation difficulty of license plate recognition is significantly less than that of comparison recognition.

[0077] In some cases, when using contrast-based recognition to identify vehicle objects within image information, a feature set is established. The specific method for establishing the feature set is as follows:

[0078] S201, when the vehicle object arrives at the entry position, obtain the first reference image of the vehicle object;

[0079] S202, establish a first reference line on the first reference image and establish feature points on the first reference line;

[0080] S203, create a feature set belonging to the first reference line based on the first reference line. The feature set includes the starting point position of the first reference line, the ending point position of the first reference line, the length of the first reference line, the angle of the first reference line, the position of the feature points, and the distance between the feature points.

[0081] Specifically, in steps S201 to S203, a first reference image of the vehicle object is first acquired when the vehicle object arrives at the entry position, such as... Figure 5 As shown, at the entrance of the parking lot, the vehicle is moving slowly and needs to stop for a period of time to enter the parking lot. At this point, a high-quality first reference image can be obtained.

[0082] After obtaining the first reference image, a first reference line is established on the first reference image. Figure 6 (as shown) and establish feature points on the first reference line ( Figure 6 (As shown), then create a feature set belonging to the first reference line based on the first reference line. The feature set includes the following:

[0083] The starting point of the first reference line, the ending point of the first reference line, the length of the first reference line, the angle of the first reference line, the position of the feature point, and the distance between the feature points.

[0084] Furthermore, the following restrictions were imposed on the first reference line, such as... Figure 7 As shown:

[0085] There are multiple first reference lines;

[0086] There are at least two parallel first reference lines and at least two intersecting first reference lines;

[0087] There are at least two parallel first reference lines. The starting points of the two first reference lines are located on the same sub-object of the vehicle object, and the ending points of the two first reference lines are located on the same sub-object of the vehicle object.

[0088] The main purpose of the above restrictions is to improve the success rate of recognition, because there are cases where vehicles are partially similar. However, when the number of first reference lines is increased, the recognition errors caused by the above situations can be avoided.

[0089] The specific steps for identification using contrast recognition are as follows:

[0090] S301, Obtain a second reference image of the vehicle object on the guide path that matches the vehicle object;

[0091] S302, Obtain local feature information of the vehicle object on the second reference image;

[0092] S303, use local feature information to determine the relative position of image sensor 1 and vehicle object corresponding to the second reference image;

[0093] S304, Select a suitable image sensor 1 according to the relative position and acquire the second reference image again;

[0094] S305, extract the first reference line and the feature set of the first reference line from the second reference image, and denote it as the contrast feature set;

[0095] S306, compare the feature set and the feature set to obtain the comparison result;

[0096] S307, determine the vehicle object based on the comparison results.

[0097] In steps S301 to S307, a second reference image of the vehicle object is first obtained on the guide path matching the vehicle object, and then local feature information of the vehicle object is obtained on the second reference image. The local feature information here refers to the parts belonging to the vehicle object, generally including parts with obvious features such as license plates, windows, rearview mirrors and tires.

[0098] Then, the relative position of the image sensor 1 corresponding to the second reference image and the vehicle object is determined using local feature information. Specifically, the relative position of the image sensor 1 corresponding to the second reference image and the vehicle object is determined based on the relative positional relationship between local feature information.

[0099] As mentioned earlier, when acquiring the first reference image, the relative position of the image sensor 1 and the vehicle object is known. At this time, the relative positional relationship (such as distance) between local feature information is also determined. By comparison, the relative position of the image sensor 1 and the vehicle object corresponding to the second reference image can be determined.

[0100] Next, a suitable image sensor 1 is selected based on the relative position, and a second reference image is acquired again. The purpose is to make the generation environment of the second reference image as consistent as possible with the generation environment of the first reference image. Here, the generation environment mainly refers to the relative position of the image sensor 1 and the vehicle object.

[0101] Then, the first reference line and its feature set are extracted from the second reference image and denoted as the comparison feature set. Finally, the comparison feature set and the feature set are compared to obtain the comparison result. When the comparison result meets the requirements, the identity of the vehicle object can be determined.

[0102] When comparing feature sets, corresponding identical elements (first reference line start point position, first reference line end point position, first reference line length, first reference line angle, feature point position, and distance between feature points) are compared. They must all be identical, have the same proportion, or have differences (distance difference, numerical difference) within the allowable range. Specific comparison rules are not restricted here.

[0103] In some examples, extracting the first reference line and its feature set from the second reference image includes the following steps:

[0104] S401, Obtain features belonging to the vehicle object on the first reference image. The features belonging to the vehicle object include the starting point position of the first reference line and / or the ending point position of the first reference line.

[0105] S402, extract the feature contour belonging to the vehicle object and move the feature contour to the corresponding position on the second reference image;

[0106] S403, adjust the state of the feature contour on the second reference image so that the overlap between the feature contour and the contour at the corresponding position on the second reference image meets the requirements;

[0107] S404, Use the contour at the corresponding position on the second reference image to establish a second reference line;

[0108] S405, Use the second reference line to generate a set of contrast features for the second reference line.

[0109] In steps S401 to S405, features belonging to the vehicle object are first obtained on the first reference image. The features belonging to the vehicle object include the starting point position of the first reference line and / or the ending point position of the first reference line. Then, the feature contour belonging to the vehicle object is extracted and the feature contour is moved to the corresponding position on the second reference image. The purpose of adjusting the state of the feature contour on the second reference image is to make the overlap between the feature contour and the contour at the corresponding position on the second reference image meet the requirements.

[0110] Because only when the overlap meets the requirements can the comparison in subsequent content be more accurate.

[0111] In some possible implementations, adjusting the state of the feature contour on the second reference image includes zooming in, zooming out, and rotating.

[0112] The method for generating the contrast feature set of the second reference line using the second reference line is as follows:

[0113] S501, using the pixels on the second reference line to generate a pixel difference curve or a quadratic pixel difference curve;

[0114] S502, obtain the curvature change points of the pixel difference curve or the quadratic curve of pixel difference, and record the curvature change points as the feature point reference position and the reference distance between feature points;

[0115] S503, Match the feature point reference position with the feature point position on the first reference line based on the distance between the feature point reference position and the starting point and / or the ending point of the second reference line, and obtain the matching result;

[0116] S504 generates the corresponding reference distance between feature points based on the matching results.

[0117] In step S501, the pixel difference curve or pixel difference quadratic curve is first generated from the pixels on the second reference line. The choice between the pixel difference curve and the pixel difference quadratic curve depends on the first reference line. In other words, the first reference line and the second reference line use the pixel difference curve or the pixel difference quadratic curve simultaneously.

[0118] The method of generating pixel difference curves from pixels on the second reference line is to determine a direction, such as from left to right, and then calculate the difference between adjacent pixels on the second reference line. Then, the position of the difference queue is used as the x-axis and the difference is used as the y-axis to generate discrete points in the coordinate system, and then these discrete points are connected sequentially.

[0119] The method for generating the quadratic curve of pixel difference is to first determine a direction, such as from left to right, then calculate the difference between adjacent differences, which is recorded as the quadratic difference. Then, using the queue position of the quadratic difference as the horizontal axis and the quadratic difference as the vertical axis, discrete points are generated in the coordinate system, and then these discrete points are connected sequentially.

[0120] A curvature change point refers to a point on the pixel difference curve or pixel difference quadratic curve where the curvature suddenly changes.

[0121] The specific method for matching the feature point positions with the first reference line and obtaining the matching result is to overlap the two pixel difference curves or the quadratic pixel difference curves, and then determine the feature point positions and the distance between feature points based on the overlap result.

[0122] This is because feature points may be incomplete when they are acquired, so they cannot be matched by sorting. Therefore, this application uses the method of coinciding two pixel difference curves or pixel difference quadratic curves to perform matching. After the matching is completed, the position of the feature points can be clearly determined, and then the corresponding reference distance between feature points can be generated based on the matching result.

[0123] The embodiments described in this specific implementation are preferred embodiments of this application and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A regional property intelligent management system, characterized in that, include: Multiple image sensors are set up within an enclosed area; the image sensors are used to acquire image information within the coverage area. Multiple guides are set up within an enclosed area. The guides are used to generate guidance information based on the vehicle object. The control system communicates data with the image sensor and the guide. The control system is used to generate a guide path that matches the vehicle object, tracks the vehicle object based on the image information fed back by the image sensor, and sends guide information to the vehicle object using the guide. The control system tracks vehicle objects based on image information fed back from the image sensor, including: The image sensor is invoked along the guide path, and its orientation is adjusted according to the guide path. Obtain image information fed back by the invoked image sensor; Identify vehicle objects included in the image information; Determine the position of the vehicle object on the guide path based on the position of the vehicle object, and record it as the current position of the vehicle object; Select a guide based on the current location of the vehicle object and send guidance information to the selected guide; The vehicle objects included in the image information are identified using license plate recognition and comparison recognition methods. When the lighting conditions meet the requirements, license plate recognition is used for identification; when the lighting conditions do not meet the requirements, comparison recognition is used for identification. Identifying vehicle objects in image information using a contrast-based recognition method involves establishing a feature set, which includes: Acquire the first reference image of the vehicle object when it reaches the entry position; Establish a first reference line on the first reference image and establish feature points on the first reference line; A feature set belonging to the first reference line is created based on the first reference line. The feature set includes the starting point position of the first reference line, the ending point position of the first reference line, the length of the first reference line, the angle of the first reference line, the position of the feature points, and the distance between the feature points. Identification using contrast-based recognition includes: Obtain a second reference image of the vehicle object along the guide path that matches the vehicle object; Local feature information of the vehicle object is obtained from the second reference image; The relative position of the image sensor and the vehicle object corresponding to the second reference image is determined using local feature information. Select a suitable image sensor based on the relative position and acquire a second reference image again; Extract the first reference line and its feature set from the second reference image, and denote it as the contrast feature set. Compare and contrast the feature set and the feature set to obtain the comparison result; The vehicle is determined based on the comparison results.

2. The regional property intelligent management system according to claim 1, characterized in that, Extracting the first reference line and its feature set from the second reference image includes: Features belonging to the vehicle object are obtained on the first reference image. The features belonging to the vehicle object include the starting point position of the first reference line and / or the ending point position of the first reference line. Extract the feature contours belonging to the vehicle object and move the feature contours to the corresponding positions on the second reference image; Adjust the state of the feature contour on the second reference image so that the overlap between the feature contour and the contour at the corresponding position on the second reference image meets the requirements; A second reference line is established using the contour at the corresponding position on the second reference image; Use the second reference line to generate a set of contrast features for the second reference line.

3. The regional property intelligent management system according to claim 2, characterized in that, Adjusting the state of the feature contour on the second reference image includes zooming in, zooming out, and rotating.

4. The regional property intelligent management system according to claim 2, characterized in that, The set of contrast features used to generate the second reference line includes: Use the pixels on the second reference line to generate a pixel difference curve or a quadratic pixel difference curve; Obtain the curvature change points of the pixel difference curve or the quadratic curve of pixel difference, and record the curvature change points as the reference positions of feature points and the reference distances between feature points; The matching result is obtained by matching the feature point reference position with the feature point position on the first reference line based on the distance between the feature point reference position and the starting point and / or the ending point of the second reference line; Generate the corresponding reference distance between feature points based on the matching results.

5. The regional property intelligent management system according to claim 1, characterized in that, There are multiple first reference lines.

6. The regional property intelligent management system according to claim 5, characterized in that, There are at least two parallel first reference lines and at least two intersecting first reference lines.

7. The regional property intelligent management system according to claim 5, characterized in that, There are at least two parallel first reference lines. The starting points of the two first reference lines are located on the same sub-object of the vehicle object, and the ending points of the two first reference lines are located on the same sub-object of the vehicle object.

Images