Target object detection method and system applied to vehicle

Abstract

The invention provides a target object detection method applied to a vehicle, and the method comprises the steps: building an object detection model, and building a corresponding relation between the signal features of a wireless signal and an object type; periodically transmitting a detection wireless signal, and receiving a detection echo signal reflected by a detected object; and importing the detection echo signal into the object detection model, analyzing the signal characteristics of the detection echo signal, determining the object type of the detected object, and screening the target object. Compared with a radar, the coverage range of the wireless router is wider, and obstacle detection without dead angles can be formed. In addition, a dead angle area at the bottom of the vehicle is also a wireless router detection range, so that rolling compaction can be avoided. Different from common millimeter wave radar detection, wireless signals with high penetrating power are applied to automatic driving, the capacity of reconnaissance on pedestrians behind obstacles is improved, the possibility that the pedestrians jump out suddenly is predicted more effectively, and the driving safety is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of wireless routing technology, and in particular to a target object detection method and system applied to a vehicle. Background Art

[0002] During the driving process and at traffic lights, pedestrians or non-motor vehicles may be blocked by obstacles or vehicles. Since the line of sight is blocked by obstacles, pedestrians and non-motor vehicles may suddenly appear from behind obstacles, commonly known as "ghost heads". The suddenness of the incident will bring huge safety hazards to the driver's normal driving. At present, the millimeter-wave radars, cameras, and lidars installed in vehicles are difficult to detect targets behind obstacles and issue corresponding alarms, and the range and accuracy of radar measurement are extremely limited, making it difficult to provide drivers with more accurate and timely warning information. When the driver encounters this situation, it is difficult to brake in time, which leads to accidents. At the same time, when the driver is in the car, the bottom of the car is also a visual blind spot for the driver, and it is difficult for the driver to observe the situation under the car, so there is a risk of crushing. Summary of the invention

[0003] The present invention provides a target object detection method applied to a vehicle. Different from the millimeter wave radar detection in the prior art, the present invention applies wireless signals with strong penetrating power to automatic driving, which can effectively predict the possibility of a pedestrian suddenly jumping out and increase driving safety.

[0004] Specifically, the object of the present invention is to provide a target object detection method applied to a vehicle, comprising:

[0005] Establish an object detection model and establish the correspondence between signal features and object types;

[0006] Periodically transmit detection wireless signals and receive detection echo signals reflected by the detected object;

[0007] The detection echo signal is introduced into the object detection model, the signal characteristics of the detection echo signal are analyzed, the object type of the detected object is determined, and the target object is screened.

[0008] Preferably, the establishing of the object detection model comprises:

[0009] Periodically transmitting a first wireless signal, and receiving a first echo signal formed after the first wireless signal is reflected by a first object;

[0010] analyzing the first echo signal to extract a first signal feature of the first echo signal;

[0011] Establishing and storing a correspondence between the first signal feature and the object type of the first object;

[0012] The object detection model is trained by a deep learning algorithm.

[0013] Preferably, the object types include in-vehicle equipment, pedestrians, non-motor vehicles, motor vehicles and road facilities; and the target objects are pedestrians and non-motor vehicles.

[0014] Preferably, a detection wireless signal is periodically transmitted and a detection echo signal is received;

[0015] The detection echo signal is introduced into the object detection model, the signal characteristics of the detection echo signal are analyzed, the signal characteristics of the target object are screened out from the signal characteristics of the detection echo signal, and the object type of the target object is determined.

[0016] Preferably, the signal characteristics include signal strength, frequency and phase; and the wireless signal is a Wifi signal.

[0017] Preferably, a detection wireless signal is periodically transmitted and a detection echo signal is received;

[0018] Determine the object type of the detected object, classify and store the detection echo signals according to the object type of the detected object, and screen the target object;

[0019] Analyzing changes in the signal characteristics of the target object in each cycle according to the object type of the target object to determine the motion characteristics of the target object;

[0020] A vehicle alarm decision is determined based on the motion characteristics of the target object.

[0021] Preferably, the motion characteristics of the target object include speed, acceleration, distance, motion direction and motion trajectory; based on the motion characteristics of the target object, the braking distance of the vehicle is calculated, and the alarm decision of the vehicle is determined based on the braking distance.

[0022] Preferably, when the braking distance is within a first threshold range, the host vehicle issues a collision reminder;

[0023] When the braking distance is within a second threshold range, the host vehicle issues a collision reminder and warning;

[0024] When the braking distance is within a third threshold range, the host vehicle issues a collision reminder and warning, and triggers automatic braking;

[0025] The first threshold range, the second threshold range and the third threshold range extend outward in sequence from the vehicle as the center.

[0026] Another aspect of the present invention provides a target object detection system applied to a vehicle, which uses any of the above-mentioned target object detection methods applied to a vehicle.

[0027] Preferably, it includes: an information collection module, an analysis module and an alarm module;

[0028] The information acquisition module is used to establish an object detection model, establish a corresponding relationship between signal characteristics and object types, and periodically transmit detection wireless signals and receive detection echo signals;

[0029] The analysis module is used to import the detection echo signal into the object detection model, analyze the signal characteristics of the detection echo signal, and determine the type of the target object;

[0030] The alarm module is used to determine a vehicle alarm decision according to the motion characteristics of the target object.

[0031] Compared with the prior art, the above technical solution has the following beneficial effects:

[0032] 1. Compared with radar, the wireless routing in the present invention has a wider coverage range, and can include the 360-degree detection range of the vehicle, forming an obstacle detection without blind spots. In addition, the blind spot area under the vehicle is also within the detection range of the wireless routing, which can avoid running over. Different from the usual millimeter-wave radar detection, the use of wireless signals with strong penetrating power in autonomous driving improves the ability to detect pedestrians behind obstacles, more effectively predicts the possibility of pedestrians suddenly jumping out, and effectively increases driving safety.

[0033] 2. In the existing technology, due to the limitations of a series of factors such as radar measurement range and accuracy, in order to accurately identify objects, the number of vehicle-mounted radars that need to be installed in a vehicle is 3-4. However, wireless routers have the characteristics of wide coverage and variable radiation direction. One router can replace 3-4 vehicle-mounted radars, reducing firmware and costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 The figure is a flow chart of a target object detection method applied to a vehicle according to an embodiment of the present invention.

[0035] Figure 2 The figure is a flow chart of a target object detection method applied to a vehicle in another embodiment of the present invention.

[0036] Figure 3 The figure is a flow chart of a target object detection method applied to a vehicle in another embodiment of the present invention.

[0037] Figure 4 This is a target object detection system applied to a vehicle in another embodiment of the present invention. DETAILED DESCRIPTION

[0038] The advantages of the present invention are further described below in conjunction with the accompanying drawings and specific embodiments.

[0039] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0040] It should be understood that although the terms first, second, third, etc. may be used in the present disclosure to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present disclosure, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0041] Figure 1 A target object detection method applied to a vehicle in an embodiment of the present invention is characterized by comprising:

[0042] Establish an object detection model and establish the correspondence between signal features and object types;

[0043] Periodically transmit detection wireless signals and receive detection echo signals reflected by the detected object;

[0044] The detection echo signal is introduced into the object detection model, the signal characteristics of the detection echo signal are analyzed, the object type of the detected object is determined, and the target object is screened.

[0045] In this embodiment, the establishment of the object detection model specifically includes: periodically transmitting a first wireless signal, receiving a first echo signal formed after the first wireless signal is reflected by a first object; analyzing the first echo signal, extracting a first signal feature of the first echo signal; establishing and storing a correspondence between the first signal feature and the object type of the first object; and training the object detection model through a deep learning algorithm. Among them, the deep learning algorithm can be any commonly used deep learning algorithm in the field. In this embodiment, considering that the target object may include pedestrians, or the most important object to be identified includes the human body, the deep learning algorithm in this embodiment can use Open pose to judge the posture of the human body. Open pose is an open source library based on convolutional neural networks and supervised learning and written in the caffe framework, which can realize the recognition of human bodies and postures.

[0046] Using Open Pose for deep learning algorithms first requires the construction of the scene and the marking of objects in the scene. In this embodiment, according to the different characteristics of the echo signals of different objects, the echo signal reflected by the human body is identified and filtered and separated, and then the model is trained based on the large number of samples obtained to achieve the highest possible accuracy.

[0047] In this embodiment, the strong penetrating power of wireless signals is used, and the different signal strengths reflected by different objects are used to assist the high-precision discrimination of deep learning to identify the type of the target object. In addition, in this embodiment, the corresponding relationship between the signal of the object and the object type can also be directly established.

[0048] It can be understood that the object types include in-vehicle equipment, pedestrians, non-motor vehicles, motor vehicles and road facilities; the target objects are pedestrians and non-motor vehicles.

[0049] Therefore, when building the model, the wireless router is first used to collect a large number of reflection signals from common objects on the road and in-vehicle equipment, and then marked to form training samples. After obtaining enough samples, the model is trained to achieve the effect of object recognition. In actual use, the wireless router will first send a wireless signal, and after receiving the reflection signal, it will import the signal into the trained model to distinguish the type of object.

[0050] In another embodiment of the present invention, the bottom of the vehicle is also included in the detection range to prevent the vehicle from running over objects, which can further enhance the 360° no-dead-angle detection range of the wireless signal.

[0051] Specifically, Figure 2As shown, in this embodiment, establishing an object detection model includes: sending a wireless signal to an in-vehicle device, receiving an echo signal of the in-vehicle device, extracting a signal feature of the echo signal, and storing the signal feature of the in-vehicle device in the object detection model;

[0052] Subsequently, the wireless router of the vehicle periodically transmits a first wireless signal and receives a detection echo signal reflected by the detected object. The first wireless signal includes a wireless signal sent toward the bottom of the vehicle.

[0053] The detection echo signal is analyzed, the signal features of the detection echo signal are extracted, and the signal features of the in-vehicle device signal are screened out; and whether there is a target object under the vehicle is determined according to the screened signal features.

[0054] Furthermore, if there is a target object under the vehicle, its specific type and motion characteristics as well as subsequent alarm decisions can be determined based on the filtered signal characteristics.

[0055] Specifically, Figure 3 As shown, it is a schematic diagram of a target object detection method applied to a vehicle in another embodiment of the present invention: the wireless router of the vehicle periodically transmits a detection wireless signal and receives a detection echo signal;

[0056] Determine the object type of the detected object, classify and store the detection echo signals according to the object type of the detected object, and screen the target object;

[0057] According to the object type of the target object, analyzing the changes of the signal characteristics of the target object in each cycle to determine the motion characteristics of the target object; the motion characteristics of the target object include speed, acceleration, distance, motion direction and motion trajectory;

[0058] The braking distance of the vehicle is calculated according to the motion characteristics of the target object, and the alarm decision of the vehicle is determined according to the braking distance.

[0059] The alarm decision includes: when the braking distance is within the first threshold range, the vehicle issues a collision reminder; when the braking distance is within the second threshold range, the vehicle issues a collision reminder and warning; when the braking distance is within the third threshold range, the vehicle issues a collision reminder, warning, and triggers automatic braking. Figure 4 As shown, the first threshold range, the second threshold range and the third threshold range extend outward in sequence from the vehicle as the center.

[0060] In this embodiment, by observing and judging multiple periodic data, the target object can be continuously positioned, and the target's motion characteristics such as motion posture, motion trajectory, speed and acceleration can be obtained. Based on the above data, the motion direction, collision possibility and estimated time of collision of the target object behind the obstacle can be judged. And the braking distance of the vehicle is calculated according to the motion characteristics. In this embodiment, the alarm level of the vehicle is divided into three levels according to the braking distance, and different alarm levels are provided respectively. When the collision possibility is lower than the critical value, the system will issue a reminder. When the collision possibility is greater than the critical value, the system will control the vehicle to brake automatically, reduce the possibility of accidents, and improve driving safety.

[0061] It should be noted that the wireless signal defined in the present invention can be a frequency such as a Wifi signal. For example, the wireless signal in the embodiment of the present invention is emitted by a vehicle-mounted wireless router, and the vehicle-mounted wireless router uses a 2.4GHz wireless router with an operating frequency of 2.402GHz-2.483Ghz.

[0062] Another embodiment of the present invention provides a target object detection system, including: an information collection module, an analysis module and an alarm module;

[0063] The information acquisition module is used to establish an object detection model, establish a corresponding relationship between signal characteristics and object types, and periodically transmit detection wireless signals and receive detection echo signals;

[0064] The analysis module is used to import the detection echo signal into the object detection model, analyze the signal characteristics of the detection echo signal, and determine the type of the target object;

[0065] The alarm module is used to determine the vehicle alarm decision according to the motion characteristics of the target object. The target object detection system in this embodiment applies any of the target object detection methods described above, and the present invention will not be described in detail here.

[0066] It should be noted that the embodiments of the present invention have better practicability and do not impose any form of limitation on the present invention. Any technician familiar with the field may use the technical content disclosed above to change or modify it into an equivalent effective embodiment. However, any modification or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A target object detection method applied to a vehicle, characterized in that: include: Establish an object detection model and establish the correspondence between the signal characteristics of wireless signals and the object types; Periodically transmit detection wireless signals and receive detection echo signals reflected by the detected object; The detection echo signal is introduced into the object detection model, the signal characteristics of the detection echo signal are analyzed, the object type of the detected object is determined, and the target object is screened.

2. The target object detection method applied to a vehicle according to claim 1, characterized in that: The object detection model is established including: Periodically transmitting a first wireless signal, and receiving a first echo signal formed after the first wireless signal is reflected by a first object; analyzing the first echo signal to extract a first signal feature of the first echo signal; Establishing and storing a correspondence between the first signal feature and the object type of the first object; The object detection model is trained by a deep learning algorithm.

3. The target object detection method applied to a vehicle according to claim 2, characterized in that: The object types include in-vehicle equipment, pedestrians, non-motor vehicles, motor vehicles and road facilities; The target objects are pedestrians and non-motor vehicles.

4. The target object detection method applied to a vehicle according to claim 3, characterized in that: Periodically transmit detection wireless signals and receive detection echo signals; The detection echo signal is introduced into the object detection model, the signal characteristics of the detection echo signal are analyzed, the signal characteristics of the target object are screened out from the signal characteristics of the detection echo signal, and the object type of the target object is determined.

5. The target object detection method applied to a vehicle according to any one of claims 1 to 4, characterized in that: The signal characteristics include signal strength, frequency and phase; The wireless signal is a Wifi signal.

6. The target object detection method applied to a vehicle according to claim 5, characterized in that: Periodically transmit detection wireless signals and receive detection echo signals; Determine the object type of the detected object, classify and store the detection echo signals according to the object type of the detected object, and screen the target object; Analyzing changes in the signal characteristics of the target object in each cycle according to the object type of the target object to determine the motion characteristics of the target object; A vehicle alarm decision is determined based on the motion characteristics of the target object.

7. The target object detection method applied to a vehicle according to claim 5, characterized in that: The motion characteristics of the target object include speed, acceleration, distance, motion direction and motion trajectory; The braking distance of the vehicle is calculated according to the motion characteristics of the target object, and the alarm decision of the vehicle is determined according to the braking distance.

8. The target object detection method applied to a vehicle according to claim 7, characterized in that: When the braking distance is within a first threshold range, the host vehicle issues a collision reminder; When the braking distance is within a second threshold range, the host vehicle issues a collision reminder and warning; When the braking distance is within a third threshold range, the host vehicle issues a collision reminder and warning, and triggers automatic braking; The first threshold range, the second threshold range and the third threshold range extend outward in sequence from the vehicle as the center.

9. A target object detection system applied to a vehicle, characterized in that: Application: A target object detection method applied to a vehicle as described in any one of claims 1-8.

10. The target object detection system for a vehicle according to claim 9, characterized in that: include: Information collection module, analysis module and alarm module; The information acquisition module is used to establish an object detection model, establish a corresponding relationship between signal characteristics and object types, and periodically transmit detection wireless signals and receive detection echo signals; The analysis module is used to import the detection echo signal into the object detection model, analyze the signal characteristics of the detection echo signal, and determine the type of the target object; The alarm module is used to determine a vehicle alarm decision according to the motion characteristics of the target object.

Images