Method, device and electronic equipment for identifying target objects in infrared images

By identifying and deriving preset areas and edges in infrared images, the three-dimensional targets of the vehicle are generated, and the real-time and accuracy of 3D object detection in monocular infrared cameras are solved, and driving safety and convenience are improved.

CN113158765BActive Publication Date: 2024-11-19YANTAI RAYTRON TECH CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202110204459.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-02-23
Publication Date
2024-11-19
Estimated Expiration
2041-02-23

AI Technical Summary

Technical Problem

The prior art is difficult to achieve real-time and accurate 3D object detection in infrared images, especially in infrared images acquired by monocular cameras, which cannot meet the real-time and safety requirements of vehicle driving.

Method used

By identifying the preset areas in the target infrared image, determining the target edges and faces of the target object, generating three-dimensional information of the target object, using the preset points and edge detection algorithm to divide the area blocks, calculate the pixel mean and intersection distance, and derive the target object's Three-dimensional target.

Benefits of technology

It realizes clear identification of vehicle position and direction in a low visibility environment, improves driving safety, simplifies structural design, and reduces computing resource requirements.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113158765B_ABST
    Figure CN113158765B_ABST
Patent Text Reader

Abstract

The present application discloses a method, device and electronic device for identifying a target object in an infrared image. The method comprises identifying a preset area contained in a target infrared image, wherein the preset area comprises the target object; determining a target edge of the target object according to at least two preset points included in the preset area; determining a first surface corresponding to the target object according to a first intersection of the target edge with a first edge of the preset area, a second intersection with a third edge of the preset area and a first preset vertex of the first edge; determining a second surface corresponding to the target object according to the second intersection and a second edge of the preset area; and generating a three-dimensional target corresponding to the target object according to the first surface and the second surface. The present application realizes that the position and driving direction of surrounding vehicles can be more clearly displayed when the vehicle is in a low-visibility environment compared to a two-dimensional image, thereby improving the safety of vehicle driving.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of image recognition, and in particular to a method, a device and an electronic device for recognizing a target object in an infrared image. Background Art

[0002] In driving environments with poor visibility and limited field of vision, such as at night, in fog and haze, and inadequate lighting, the driver's visual distance will be significantly shortened, and the high beam will be dazzling when meeting other vehicles, which will greatly affect and threaten driving safety. The on-board infrared system using infrared imaging technology can help the driver clearly observe the surrounding vehicles and pedestrians in these adverse scenes, and can further see the areas that are not illuminated by visible light, and discover safety warning information in advance.

[0003] Thanks to the development and application of deep learning technology, the technology for 2D object detection in images has gradually matured, and the detection efficiency and accuracy have been improved to a certain extent. However, 2D object detection can only obtain two-dimensional information of the target. For vehicles and drivers in driving scenarios such as autonomous driving and / or driving environments with poor field of view, more accurate information about surrounding vehicles is required.

[0004] In the prior art, methods for 3D object detection on visible light images are mainly divided into feature-based 3D object detection and deep learning-based 3D object detection.

[0005] The feature-based 3D target detection method needs to extract a large number of local feature point features of the target from the pre-created 3D model database, and use the extracted target features to match the features of the target to be detected to achieve the goal of 3D detection. However, this method requires sufficient texture on the target to be detected to calculate the features, but infrared images lack sufficient texture features, so it is difficult to apply to infrared images.

[0006] 3D object detection based on deep learning can make comprehensive use of monocular cameras, binocular cameras, and multi-line lidar for 3D object detection. Since the monocular camera algorithm is highly efficient and low-cost, the monocular camera-based method is more widely used.

[0007] There are two main types of deep learning-based 3D object detection methods applied to monocular cameras: one is to expand 2D detection methods, and the other is to estimate 3D information first and then perform 3D detection. However, deep learning-based 3D object detection methods applied to monocular cameras require huge neural network calculations to extract the features and position information of the target. The efficiency of the algorithm cannot meet the real-time requirements of the infrared vehicle-mounted auxiliary system. In addition, the current open source 3D detection datasets are all based on data collected from visible light, and there is a lack of 3D object detection datasets that can be applied to infrared images.

[0008] Therefore, developing a 3D detection method that can be applied to infrared images captured by a monocular infrared camera to obtain the size, position and direction of the vehicle is of great significance to improving the convenience and safety of vehicle driving. Summary of the invention

[0009] In order to solve the deficiencies of the prior art, the main purpose of the present invention is to provide a method, device and electronic device for identifying a target object in an infrared image, so as to improve the convenience and safety of vehicle driving. In order to achieve the above-mentioned purpose, in a first aspect, the present invention provides a method for identifying a target object in an infrared image, the method comprising:

[0010] Identify a preset area contained in the target infrared image, wherein the preset area includes the target object;

[0011] Determining a target edge of the target object according to at least two preset points included in the preset area;

[0012] Determine a first surface corresponding to the target object according to a first intersection point of the target edge with the first edge of the preset area, a second intersection point with the third edge of the preset area, and a first preset vertex of the first edge of the preset area;

[0013] Determining a second surface corresponding to the target object according to the second intersection point and a second edge of the preset area;

[0014] A three-dimensional object corresponding to the target object is generated according to the first surface and the second surface.

[0015] In some embodiments, determining the target edge of the target object according to at least two preset points included in the preset area includes:

[0016] According to a preset edge detection method, identifying the outline of the target object contained in the preset area;

[0017] Dividing the outline into area blocks according to a preset division rule;

[0018] Acquire a first preset area block and a second preset area block located at a first preset position from among all the area blocks;

[0019] Determine a region block in which the pixel mean value in the first preset region block and the second preset region block meets a preset condition as a target region block;

[0020] Points at second preset positions in the target area block are determined as the at least two preset points.

[0021] In some embodiments, determining the first surface corresponding to the target object according to the first intersection of the target edge and the first edge of the preset area, the second intersection with the third edge of the preset area, and the first preset vertex of the first edge of the preset area includes:

[0022] A third intersection is determined based on the first intersection, the second intersection and the first preset vertex, the difference between the length of the line between the third intersection and the second intersection and the distance between the first preset vertex and the first intersection does not exceed a first preset threshold, and the angle between the line and the first edge does not exceed a second preset threshold.

[0023] In some embodiments, the second edge includes a second preset vertex and a third preset vertex, and determining the second surface corresponding to the target object according to the second intersection point and the second edge of the preset area includes:

[0024] A second surface corresponding to the target object is determined according to the third intersection point, the second intersection point, the second preset vertex, and the third preset vertex.

[0025] In some embodiments, the size of the preset area is not less than a preset image size threshold.

[0026] In some embodiments, the target object is a vehicle, and the method includes:

[0027] According to the lane line and the target edge contained in the infrared image to be identified, the displacement direction of the vehicle to be identified relative to the lane line is determined.

[0028] In some embodiments, the second predetermined vertex is formed by the intersection of the third edge and the second edge, and the method includes:

[0029] Determine the length of the target object according to the distance between the first intersection point and the second intersection point in the generated three-dimensional target; and / or

[0030] Determine the height of the target object according to the distance between the first preset vertex and the first intersection in the generated three-dimensional target; and / or

[0031] The width of the target object is determined according to the distance between the second intersection point and the second preset vertex in the generated three-dimensional target.

[0032] In some embodiments, the shape of the preset area is a quadrilateral.

[0033] In a second aspect, the present application provides a device for identifying a vehicle in an infrared image, the device comprising:

[0034] An identification module, used for identifying a preset area contained in a target infrared image, wherein the preset area includes the target object;

[0035] A determination module, configured to determine a target edge of the target object according to at least two preset points included in the preset area;

[0036] The judgment module is further used to determine the first surface corresponding to the target object according to the first intersection point of the target edge with the first edge of the preset area, the second intersection point with the third edge of the preset area, and the first preset vertex of the first edge of the preset area;

[0037] The judgment module is further used to determine the second surface corresponding to the target object according to the second intersection point and the second edge of the preset area;

[0038] A generating module is used to generate a three-dimensional target corresponding to the target object according to the first surface and the second surface.

[0039] In a third aspect, the present application provides an electronic device, including:

[0040] one or more processors; and

[0041] A memory associated with the one or more processors, the memory being used to store program instructions, wherein when the program instructions are read and executed by the one or more processors, the following operations are performed:

[0042] Identify a preset area contained in the target infrared image, wherein the preset area includes the target object;

[0043] Determining a target edge of the target object according to at least two preset points included in the preset area;

[0044] Determine a first surface corresponding to the target object according to a first intersection point of the target edge with the first edge of the preset area, a second intersection point with the third edge of the preset area, and a first preset vertex of the first edge of the preset area;

[0045] Determining a second surface corresponding to the target object according to the second intersection point and a second edge of the preset area;

[0046] A three-dimensional object corresponding to the target object is generated according to the first surface and the second surface.

[0047] The beneficial effects achieved by the present invention are:

[0048] The present application provides a method for identifying a target object in an infrared image, comprising identifying a preset area contained in a target infrared image, wherein the preset area includes the target object; determining a target edge of the target object according to at least two preset points included in the preset area; determining a first surface corresponding to the target object according to a first intersection of the target edge with a first edge of the preset area, a second intersection with a third edge of the preset area, and a first preset vertex of the first edge of the preset area; determining a second surface corresponding to the target object according to the second intersection and a second edge of the preset area; and generating a three-dimensional target corresponding to the target object according to the first surface and the second surface. The present application identifies a preset area containing a target object in an infrared image collected by a monocular infrared camera through a detection algorithm, wherein the target object may include objects such as vehicles, and further generates a three-dimensional target of the three-dimensional target object according to the two-dimensional preset area. Compared with a two-dimensional image, the position and driving direction of surrounding vehicles can be more clearly displayed when the vehicle is in a low-visibility environment, thereby improving the safety of vehicle driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0050] Figure 1 It is a schematic diagram of vehicle area identification provided by an embodiment of the present application;

[0051] Figure 2 is a schematic diagram of a target area identification process provided by an embodiment of the present application;

[0052] Figure 3 is a schematic diagram of a three-dimensional target recognition process provided by an embodiment of the present application;

[0053] Figure 4 is a schematic diagram of a three-dimensional target provided in an embodiment of the present application;

[0054] Figure 5 It is a diagram of the vehicle system architecture provided by an embodiment of the present application;

[0055] Figure 6 is a flow chart of the method provided in the embodiment of the present application;

[0056] Figure 7 is a structural diagram of a device provided in an embodiment of the present application;

[0057] Figure 8 This is an electronic device architecture diagram provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0059] As described in the background art, when a vehicle is in a driving scenario with poor visibility and limited field of vision, the safe driving of the vehicle cannot be guaranteed by the driver's naked eye observation of road conditions alone.

[0060] To solve the above problems, the present application proposes a method for identifying target objects in infrared images, which can extract three-dimensional targets such as vehicles from infrared images captured by a monocular infrared camera.

[0061] Figure 5 The system architecture diagram of the vehicle-mounted infrared system provided by the present application is shown. The vehicle-mounted infrared system includes a vehicle-mounted infrared camera and an ECU processing unit. The ECU processing unit can obtain the infrared image collected by the infrared camera, use the intelligent algorithm module to collect the corresponding 3D target according to the infrared image, and then use the display module to visualize the generated 3D target to the user through the vehicle display screen, or transmit the data to the vehicle CAN bus and the car power system through the control module and / or the sensor module, so that the vehicle automatic driving system and other systems can obtain the corresponding 3D target to make vehicle driving decisions. The vehicle-mounted infrared camera can be of the Xsafe-II M series, and the ECU processing unit can be built based on the Ambarella CV25 chip.

[0062] Embodiment 1

[0063] Specifically, the process of applying the target object recognition method in the infrared image provided by the present application to extract the vehicle 3D target includes:

[0064] Step 1: Identify and extract all or part of the vehicle area contained in the infrared image to be identified according to a preset target detection algorithm;

[0065] The infrared image to be identified can be obtained based on an infrared picture or infrared video collected by an infrared monocular camera. When the infrared image to be identified is an infrared video, each video frame contained in the infrared video can be used as the infrared image to be identified for 3D target identification. The infrared monocular camera can be pre-installed on the vehicle to realize real-time collection of road condition information.

[0066] The target detection algorithm can be any target detection algorithm, such as HOG+SVM, ACF+Adaboost based on machine learning detection algorithms and RCNN, Faster RCNN, YOLO series, SSD, mobilenet_yolo, mobilenet_SSD based on deep learning detection algorithms, etc. Preferably, the mobilenet-yolov3 algorithm based on deep learning algorithm can be used, the backbone of which adopts a lightweight mobilenet structure, which can be used to extract feature information of the target.

[0067] The target detection algorithm can extract the vehicle area contained in the infrared image to be identified, such as Figure 1 As shown, the image within the preset image frame is the extracted vehicle area.

[0068] Preferably, since it is difficult to display the 3D information of vehicles that are far away from the vehicle in a 2D infrared image, and it is difficult to extract the information of vehicles that are far away from the vehicle and the accuracy of the generated three-dimensional target is relatively low, it is possible to select vehicles whose distance from the vehicle does not exceed a preset range for 3D target recognition. Preferably, the vehicle area whose size is not less than a preset image size threshold among all the identified vehicle areas can be extracted. Preferably, the shape of the vehicle area is a quadrilateral, preferably a rectangle.

[0069] Step 2: Identify the vehicle contours included in the vehicle area according to a preset edge detection algorithm;

[0070] The Canny operator can be used to identify the edge contour of the vehicle. Since the vehicle is brighter than the background in the infrared image, a larger threshold can be selected for canny edge calculation when selecting the canny threshold. The sobel operator, Laplacian operator, Roberts algorithm, Kirsch operator and other arbitrary edge detection operators can also be used for edge detection, and this application does not limit this.

[0071] Step 3: Divide the vehicle outline into area blocks and determine the target area block according to the preset division rules;

[0072] Preferably, Figure 2 As shown, the vehicle outline can be evenly divided into six area blocks according to the rule that the height is 1 / 2 of the vehicle outline height and the width is 1 / 3 of the vehicle ROI width.

[0073] Then, two area blocks in the lower left corner or lower right corner of all area blocks containing the vehicle outline are obtained, and the pixel means of the two areas are calculated respectively. Since the pixel mean of the area containing tires is statistically larger, it can be determined that the area block with a higher pixel mean among the two area blocks contains more tires and is the target area.

[0074] Step 4: determining the target edge of the vehicle according to at least two preset points included in the target area block;

[0075] The at least two preset points include the intersection of the first tire of the vehicle and the ground and the intersection of the second tire of the vehicle and the ground, the first tire and the second tire are tires on the same side, the first tire is the front tire of all tires visible in the target area block, and the second tire is the rear tire of all tires visible in the target area block. The line between the preset points is the target edge of the vehicle.

[0076] The preset point is a point that can be used to indicate the position of the edge of a target object such as a vehicle. When the target object is a vehicle, the intersection of the first and last two tires on the same side of the vehicle and the ground can indicate the position of the edge of the side of the vehicle. Preferably, when the same side of the vehicle includes more than two tires, the number of preset points can also exceed two, that is, including the intersection of each tire on the same side and the ground. When the target object is an object other than a vehicle, the preset point can also be other points of the target object that can be used to indicate the position of the edge.

[0077] like Figure 3 As shown in FIG. 1 , when the target area block is an area block including the lower left corner, it can be determined that the left half of the vehicle area 100 includes the side of the vehicle 101, that is, the vehicle area 100 includes the left side of the vehicle 101. Figure 3 As shown in the vehicle 201 included in the vehicle area 200, when the target area block is the area block including the lower right corner, it can be determined that the right half of the vehicle area 200 includes the side of the vehicle 201, which is equivalent to the vehicle area 200 including the right side of the vehicle 201.

[0078] like Figure 3As shown, a plane rectangular coordinate system can be established with any point on the infrared image to be identified as the origin, wherein the X-axis is parallel to the lateral direction of the image, and the Y-axis is parallel to the longitudinal direction of the image. The intersection point a between the first tire 102 of the vehicle 101 and the ground is the point with the smallest ordinate among all the coordinate points with the smallest abscissa value contained in the image corresponding to the first tire; the intersection point b between the second tire 103 and the ground is the point with the lowest ordinate contained in the image corresponding to the second tire. The intersection point between the first tire of the vehicle 201 and the ground is the point with the smallest ordinate among all the coordinate points with the largest abscissa value contained in the image corresponding to the first tire, and the intersection point between the second tire and the ground is the point with the lowest ordinate contained in the image corresponding to the second tire. The positions of the intersection points of the first tire and the second tire with the ground can also be determined according to any other method, and this application does not limit this.

[0079] Step 5: determining a first surface corresponding to the vehicle according to a first intersection point of the target edge with the first edge of the vehicle region, a second intersection point with the third edge, and a first preset vertex of the third edge;

[0080] When the target area block is an area block including the lower left corner, the first edge is the left edge of the vehicle area; when the target area is an area block including the lower right corner, the first edge is the right edge of the vehicle area. The third edge is the bottom edge of the image.

[0081] like Figure 3 As shown, taking vehicle 101 as an example, the intersection point of the straight line passing through intersection a and intersection b and the third edge 104 is the second intersection point c; the intersection point of the straight line passing through intersection a and intersection b and the first edge 104 is the first intersection point d.

[0082] The difference between the abscissa of the third intersection e and the abscissa of the second intersection c does not exceed the third preset threshold, and the difference between the ordinate and the sum of the ordinate of the second intersection c and the height of the vehicle profile does not exceed the first preset threshold. The third preset threshold can be determined based on the value range of the abscissa of the third intersection e when the angle between the connecting line and the first edge does not exceed the second preset threshold.

[0083] Preferably, the abscissa of the third intersection e is the same as the abscissa of the second intersection c, and the ordinate is the sum of the ordinate of the second intersection c and the height of the vehicle profile. Specifically, the height of the vehicle profile can be determined according to the length of the distance between the first preset vertex and the first intersection.

[0084] According to the first preset vertex f, the second intersection c, the first intersection d and the third intersection e, the first face, i.e., the side face, corresponding to the vehicle to be identified can be determined. The first edge includes the first preset vertex f. When the target edge is on the left side of the vehicle, i.e., the first edge is the left edge of the vehicle area, the first preset vertex f is the upper left vertex of the vehicle area; when the target edge is on the right side of the vehicle, i.e., the first edge is the right edge of the vehicle area, the first preset vertex f is the upper right vertex of the vehicle area.

[0085] Step 6: Determine a second surface corresponding to the vehicle according to the second intersection point and the second edge of the vehicle area;

[0086] like Figure 3 As shown, the second edge includes the second preset vertex g and the third preset vertex h. The second side, i.e., the back side, corresponding to the vehicle to be identified can be determined according to the second intersection c, the third intersection e, the second preset vertex g, and the third preset vertex h. When the target edge is on the left side of the vehicle, the second preset vertex is the upper right vertex of the vehicle area, and the third preset vertex is the lower right vertex of the image. When the target edge is on the right side of the vehicle, the second preset vertex is the lower left vertex of the vehicle area, and the third preset vertex is the lower left vertex of the image.

[0087] Step 7: Determine the three-dimensional target corresponding to the vehicle according to the first surface and the second surface;

[0088] like Figure 4 As shown, a three-dimensional target corresponding to the vehicle can be generated based on the first surface and the second surface.

[0089] The length, width and height of the vehicle can be determined based on the generated three-dimensional target. The specific process includes:

[0090] Determining the length of the vehicle based on the distance between the first intersection point and the second intersection point in the generated three-dimensional object; and / or

[0091] Determine the height of the vehicle according to the distance between the first preset vertex and the first intersection in the generated three-dimensional object; and / or

[0092] The width of the vehicle is determined according to the distance between the second intersection point and the second preset vertex in the generated three-dimensional object.

[0093] When the image to be identified contains lane lines, the displacement direction of the vehicle relative to the lane lines can be determined based on the angle between the lane lines and the target edge.

[0094] Based on the method disclosed in this application, this application realizes 3D target detection at night, in dust, haze and other weather conditions based on infrared images, which can make up for the defects of visible light and improve driving safety in the above scenarios. This application uses target characteristics to perform 3D information reasoning based on 2D target detection and edge detection. Compared with 3D targets based on deep learning, it requires less computing power and has better real-time performance; compared with the multi-sensor detection method, it only requires one infrared detector, which is simple in structure and saves costs.

[0095] Embodiment 2

[0096] Corresponding to the above embodiment, Figure 6 As shown, the present application provides a method for identifying a target object in an infrared image, the method comprising:

[0097] 610. Identify a preset area contained in a target infrared image, where the preset area includes the target object;

[0098] 620. Determine a target edge of the target object according to at least two preset points included in the preset area;

[0099] Preferably, determining the target edge of the target object according to at least two preset points included in the preset area comprises:

[0100] 621. Identify the outline of the target object contained in the preset area according to a preset edge detection method;

[0101] 622. Divide the outline into area blocks according to a preset division rule;

[0102] 623. Obtain a first preset area block and a second preset area block located at a first preset position from among all the area blocks;

[0103] 624. Determine a region block in which the pixel mean values ​​of the first preset region block and the second preset region block meet a preset condition as a target region block;

[0104] 625. Determine points in the target area block that are located at second preset positions as the at least two preset points.

[0105] 630. Determine a first surface corresponding to the target object according to a first intersection point of the target edge with a first edge of the preset area, a second intersection point of the target edge with a third edge of the preset area, and a first preset vertex of the first edge;

[0106] Preferably, determining the first surface corresponding to the target object according to a first intersection point of the target edge with the first edge of the preset area, a second intersection point with the third edge of the preset area, and a first preset vertex of the first edge comprises:

[0107] 631. Determine a third intersection point according to the first intersection point, the second intersection point and the first preset vertex, wherein the difference between the length of a line between the third intersection point and the second intersection point and the distance between the first preset vertex and the first intersection point does not exceed a first preset threshold, and an angle between the line and the first edge does not exceed a second preset threshold;

[0108] 632. Determine a first surface corresponding to the target object according to the third intersection point, the second intersection point, the first intersection point and the first preset vertex.

[0109] 640. Determine a second surface corresponding to the target object according to the second intersection point and a second edge of the preset area;

[0110] Preferably, the second edge includes a second preset vertex and a third preset vertex, and determining the second surface corresponding to the target object according to the second intersection point and the second edge of the preset area includes:

[0111] 641. Determine a second surface corresponding to the target object according to the third intersection point, the second intersection point, the second preset vertex, and the third preset vertex.

[0112] 650. Generate a three-dimensional target corresponding to the target object according to the first surface and the second surface.

[0113] Preferably, the size of the preset area is not less than a preset image size threshold.

[0114] Preferably, the target object is a vehicle, and the method comprises:

[0115] 660. Determine the displacement direction of the vehicle to be identified relative to the lane line according to the lane line and the target edge included in the target infrared image.

[0116] Preferably, the second preset vertex is formed by the intersection of the third edge and the second edge, and the method comprises:

[0117] 670. Determine the length of the target object according to the distance between the first intersection point and the second intersection point in the generated three-dimensional target; and / or

[0118] 671. Determine the height of the target object according to the distance between the first preset vertex and the first intersection in the generated three-dimensional target; and / or

[0119] 672. Determine the width of the target object according to the distance between the second intersection point and the second preset vertex in the generated three-dimensional target.

[0120] Preferably, the preset area is in the shape of a quadrilateral.

[0121] Embodiment 3

[0122] Corresponding to the above embodiment, the present application provides a device for identifying a target object in an infrared image, the device group comprising:

[0123] An identification module 710 is used to identify a preset area contained in a target infrared image, wherein the preset area includes the target object;

[0124] A determination module 720, configured to determine a target edge of the target object according to at least two preset points included in the preset area;

[0125] The determination module 720 is further configured to determine a first surface corresponding to the target object according to a first intersection point of the target edge with the first edge of the preset area, a second intersection point with the third edge of the preset area, and a first preset vertex of the first edge;

[0126] The determination module 720 is further configured to determine a second surface corresponding to the target object according to the second intersection point and a second edge of the preset area;

[0127] The generating module 730 is configured to generate a three-dimensional target corresponding to the target object according to the first surface and the second surface.

[0128] Preferably, the recognition module 710 can also be used to identify the outline of the target object contained in the preset area according to a preset edge detection method; the judgment module 720 can also be used to divide the outline into area blocks according to a preset division rule; obtain a first preset area block and a second preset area block at a first preset position among all the area blocks; determine that the area blocks whose pixel means meet preset conditions in the first preset area block and the second preset area block are target area blocks; and determine that the points in the target area block at the second preset position are the at least two preset points.

[0129] Preferably, the judgment module 720 can also be used to determine a third intersection based on the first intersection, the second intersection and the first preset vertex, the difference between the length of the line between the third intersection and the second intersection and the distance between the first preset vertex and the first intersection does not exceed a first preset threshold and the angle between the line and the first edge does not exceed a second preset threshold; determine the first face corresponding to the target object based on the third intersection, the second intersection, the first intersection and the first preset vertex.

[0130] Preferably, the judgment module 720 may also be used to determine the second surface corresponding to the target object according to the third intersection point, the second intersection point, the second preset vertex and the third preset vertex.

[0131] Preferably, the size of the preset area is not less than a preset image size threshold.

[0132] Preferably, the target object is a vehicle, and the judgment module can also be used to determine the displacement direction of the vehicle to be identified relative to the lane line based on the lane line and the target edge contained in the target infrared image.

[0133] Preferably, the second preset vertex is formed by the intersection of the third edge and the second edge, and the judgment module 720 can also be used to determine the length of the target object based on the distance between the first intersection and the second intersection in the generated three-dimensional target; and / or determine the height of the target object based on the distance between the first preset vertex and the first intersection in the generated three-dimensional target; and / or determine the width of the target object based on the distance between the second intersection and the second preset vertex in the generated three-dimensional target.

[0134] Embodiment 4

[0135] Corresponding to the above method and device, an embodiment of the present application provides an electronic device, including:

[0136] One or more processors; and a memory associated with the one or more processors, the memory being used to store program instructions, the program instructions when read and executed by the one or more processors, performing the following operations:

[0137] Identify a preset area contained in the target infrared image, wherein the preset area includes the target object;

[0138] Determining a target edge of the target object according to at least two preset points included in the preset area;

[0139] Determine a first surface corresponding to the target object according to a first intersection point of the target edge with a first edge of the preset area, a second intersection point of the target edge with a third edge of the preset area, and a first preset vertex of the first edge;

[0140] Determining a second surface corresponding to the target object according to the second intersection point and a second edge of the preset area;

[0141] A three-dimensional object corresponding to the target object is generated according to the first surface and the second surface.

[0142] in, Figure 8The architecture of the electronic device is shown as an example, which may include a processor 1510, a video display adapter 1511, a disk drive 1512, an input / output interface 1513, a network interface 1514, and a memory 1520. The processor 1510, the video display adapter 1511, the disk drive 1512, the input / output interface 1513, the network interface 1514, and the memory 1520 may be communicatively connected via a communication bus 1530.

[0143] Among them, the processor 1510 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solution provided in this application.

[0144] The memory 1520 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1520 can store an operating system 1521 for controlling the operation of the electronic device 1500, and a basic input and output system (BIOS) 1522 for controlling the low-level operation of the electronic device 1500. In addition, a web browser 1523, a data storage manager 1524, and an icon font processing system 1525, etc. can also be stored. The above-mentioned icon font processing system 1525 can be an application program that specifically implements the operations of the aforementioned steps in the embodiment of the present application. In short, when the technical solution provided by the present application is implemented by software or firmware, the relevant program code is stored in the memory 1520 and is called and executed by the processor 1510. The input / output interface 1513 is used to connect the input / output module to realize information input and output. The input / output / module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input devices may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output devices may include a display, a speaker, a vibrator, an indicator light, etc.

[0145] The network interface 1514 is used to connect to a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0146] The bus 1530 comprises a pathway for transmitting information between the various components of the device (eg, the processor 1510, the video display adapter 1511, the disk drive 1512, the input / output interface 1513, the network interface 1514, and the memory 1520).

[0147] In addition, the electronic device 1500 can also obtain information on specific collection conditions from the virtual resource object collection condition information database 1541 for use in condition judgment, etc.

[0148] It should be noted that, although the above device only shows a processor 1510, a video display adapter 1511, a disk drive 1512, an input / output interface 1513, a network interface 1514, a memory 1520, a bus 1530, etc., in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include components necessary for implementing the solution of the present application, and does not necessarily include all the components shown in the figure.

[0149] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., including several instructions for enabling a computer device (which can be a personal computer, a cloud server, or a network device, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0150] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0151] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying a target object in an infrared image, characterized in that: The method comprises: Identify a preset area contained in the target infrared image, wherein the preset area includes the target object; Determining a target edge of the target object according to at least two preset points included in the preset area; Determine a first surface corresponding to the target object according to a first intersection point of the target edge with a first edge of the preset area, a second intersection point of the target edge with a third edge of the preset area, and a first preset vertex of the first edge; Determine a second surface corresponding to the target object according to the second intersection point and a second edge of the preset area, wherein the second edge includes a second preset vertex and a third preset vertex, and the second preset vertex is formed by the intersection of the third edge and the second edge; Generating a three-dimensional target corresponding to the target object according to the first surface and the second surface; Wherein, determining the target edge of the target object according to at least two preset points included in the preset area comprises: According to a preset edge detection method, identifying the outline of the target object contained in the preset area; Dividing the outline into area blocks according to a preset division rule; Acquire a first preset area block and a second preset area block located at a first preset position from among all the area blocks; Determine a region block in which the pixel mean value in the first preset region block and the second preset region block meets a preset condition as a target region block; Determining points in the target area block that are at second preset positions as the at least two preset points; The determining the first surface corresponding to the target object according to the first intersection point of the target edge with the first edge of the preset area, the second intersection point with the third edge of the preset area, and the first preset vertex of the first edge comprises: Determine a third intersection point according to the first intersection point, the second intersection point and the first preset vertex, wherein the difference between the length of a line between the third intersection point and the second intersection point and the distance between the first preset vertex and the first intersection point does not exceed a first preset threshold value, and the angle between the line and the first edge does not exceed a second preset threshold value; Determining a first surface corresponding to the target object according to the third intersection point, the second intersection point, the first intersection point and the first preset vertex; The determining, according to the second intersection point and the second edge of the preset area, the second surface corresponding to the target object comprises: A second surface corresponding to the target object is determined according to the third intersection point, the second intersection point, the second preset vertex, and the third preset vertex.

2. The method according to claim 1, characterized in that The size of the preset area is not less than a preset image size threshold.

3. The method according to claim 1, characterized in that The target object is a vehicle, and the method comprises: According to the lane line and the target edge contained in the target infrared image, the displacement direction of the vehicle to be identified relative to the lane line is determined.

4. The method according to claim 1, characterized in that: The method comprises: determining the length of the target object according to the distance between the first intersection point and the second intersection point in the generated three-dimensional target; and / or, determining the height of the target object according to the distance between the first preset vertex and the first intersection in the generated three-dimensional target; and / or, The width of the target object is determined according to the distance between the second intersection point and the second preset vertex in the generated three-dimensional target.

5. The method according to claim 1, characterized in that The shape of the preset area is a quadrilateral.

6. A recognition device for implementing the method for recognizing a target object in an infrared image according to claim 1, characterized in that: The device comprises: An identification module, used for identifying a preset area contained in a target infrared image, wherein the preset area includes the target object; A determination module, configured to determine a target edge of the target object according to at least two preset points included in the preset area; The judgment module is further used to determine the first surface corresponding to the target object according to the first intersection point of the target edge with the first edge of the preset area, the second intersection point with the third edge of the preset area and the first preset vertex of the first edge; The judgment module is further used to determine the second surface corresponding to the target object according to the second intersection point and the second edge of the preset area; A generating module is used to generate a three-dimensional target corresponding to the target object according to the first surface and the second surface.

7. An electronic device, characterized in that: include: one or more processors; as well as A memory associated with the one or more processors, the memory being used to store program instructions, wherein when the program instructions are read and executed by the one or more processors, the operations of the method for identifying target objects in infrared images as described in claim 1 are performed.

Citation Information

Patent Citations

  • Infrared CCD based foggy day identifying early-warning system and method for vehicle

    CN105206109A

  • A method and apparatus for determining a three-dimensional frame of a vehicle

    CN109829447A

  • Image synthesis method and device, and storage medium

    CN112308769A

  • Coordinate acquisition device, system and method for real-time 3D reconstruction, and stereoscopic interactive device

    WO2014044126A1