Search

CN-121973721-A - Method, system, equipment and medium for detecting children left in vehicle

CN121973721ACN 121973721 ACN121973721 ACN 121973721ACN-121973721-A

Abstract

The invention relates to the field of vehicle safety, in particular to a method, a system, equipment and a medium for detecting children left in a vehicle, which comprise the steps of acquiring the prediction of the passenger states of seats by different types of sensors for each seat; the method comprises the steps of determining initial confidence coefficient of passenger state prediction based on preset confidence coefficient of each sensor and passenger state prediction, correcting the initial confidence coefficient based on a preset correction function of each sensor to obtain correction confidence coefficient, fusing the correction confidence coefficients corresponding to all types of sensors for each seat to obtain fusion confidence coefficient corresponding to the seat, and judging whether children remain in the vehicle according to the fusion confidence coefficient corresponding to each seat. The invention remarkably reduces the negative influence on the whole judgment caused by the reduced reliability of a single sensor in a specific scene, and effectively improves the detection accuracy and the environmental robustness under different actual working conditions.

Inventors

  • CHEN JIAN

Assignees

  • 奇瑞汽车股份有限公司

Dates

Publication Date
20260505
Application Date
20260113

Claims (10)

  1. 1. A method of detecting a child left in a vehicle, the method comprising: Acquiring, for each seat, a prediction of an occupant condition of the seat by a different type of sensor; determining an initial confidence of the occupant state prediction based on the preset confidence of each sensor and the occupant state prediction; correcting the initial confidence coefficient based on a preset correction function of each sensor to obtain a correction confidence coefficient; Fusing correction confidence degrees corresponding to all types of sensors for each seat to obtain fusion confidence degrees corresponding to the seats; And judging whether children remain in the vehicle according to the fusion confidence degree corresponding to each seat.
  2. 2. The method of claim 1, wherein the occupant condition prediction comprises whether an occupant is present in the seat; and the obtaining mode of the preset confidence coefficient of the sensor comprises the following steps: And performing a supervision test on the sensor, wherein the test times exceed the preset minimum test times, the test content is whether the seat has passengers or not, and the prediction accuracy of the sensor on the passengers and the absence of the passengers is counted and used as the preset confidence.
  3. 3. The method for detecting a child left in a vehicle according to claim 1, wherein the occupant state prediction includes an age group of an occupant; and the obtaining mode of the preset confidence coefficient of the sensor comprises the following steps: And performing supervised tests on the sensor, wherein the test times exceed the preset minimum test times, the test contents are preset age groups to which the predicted passengers belong, and the prediction accuracy of the sensor for each preset age group is counted and used as the preset confidence coefficient corresponding to each age group.
  4. 4. The method for detecting a child left in a vehicle according to claim 1, wherein the obtaining means of the correction function comprises: And collecting historical detection data and corresponding real states of the sensor in a real vehicle environment, establishing a relation model between the initial confidence coefficient and the real states through a data fitting method, and determining the relation model as the preset correction function.
  5. 5. The method for detecting a child left in a vehicle according to claim 1, wherein the step of fusing the correction confidences corresponding to all types of sensors for each seat to obtain the fusion confidences corresponding to the seats comprises: calculating the fusion confidence according to the following formula : ; Wherein, the Represents the first The correction confidence corresponding to each type of sensor, Representing the number of sensor types.
  6. 6. The method for detecting children remaining in a vehicle according to claim 1, wherein the step of determining whether the children remain in the vehicle according to the fusion confidence level corresponding to each seat comprises: Comparing the fusion confidence coefficient of each seat with a preset confidence coefficient threshold value, and judging whether children exist in the seat or not; if there is a child and no adult occupant is found based on the determination of all seats, it is determined that a child is left in the vehicle.
  7. 7. The method for detecting a child left in a vehicle according to claim 1, further comprising: and when the child is judged to be left in the vehicle, automatically executing at least one preset safety protection measure, wherein the safety protection measure comprises triggering an audible and visual alarm device,/or sending an alarm notification to preset terminal equipment,/or starting an in-vehicle environment regulating system, and/or making an alarm call.
  8. 8. A system for detecting a child left in a vehicle, comprising: An acquisition module for acquiring, for each seat, a prediction of an occupant state of the seat by different types of sensors; a determining module configured to determine an initial confidence of the occupant state prediction based on a preset confidence of each sensor and the occupant state prediction; the correction module is used for correcting the initial confidence coefficient based on a preset correction function of each sensor to obtain a correction confidence coefficient; The fusion module is used for fusing the correction confidence degrees corresponding to all types of sensors for each seat to obtain the fusion confidence degrees corresponding to the seats; And the judging module is used for judging whether children remain in the vehicle according to the fusion confidence degree corresponding to each seat.
  9. 9. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, carries out the steps of the method for detecting a legacy child in a vehicle according to any one of claims 1 to 7.
  10. 10. A storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the method of detecting a child left in a vehicle as claimed in any one of claims 1 to 7.

Description

Method, system, equipment and medium for detecting children left in vehicle Technical Field The invention relates to the field of vehicle safety, in particular to a method, a system, equipment and a medium for detecting children left in a vehicle. Background The child carry-over detection is based on an in-vehicle monitoring technology of electronic elements such as sensors, and is used for identifying whether a child is still left in the vehicle after flameout and locking of the vehicle and intervening when the child is confirmed to be left, so that the occurrence of child heat shooting, asphyxia and even death caused by sudden changes in environments such as high temperature, hypoxia and the like is avoided. The child carry-over detection can save lives, reduce occurrence of household tragedy, promote evolution of automobile safety from passive protection to active care, and embody the necessary trend of humanization of technology in the intelligent automobile age. It is counted that in an external environment at 35 ℃, the temperature in the closed compartment can rise to 65 ℃ within 15 minutes, and children can suffer irreversible organic damage in the environment only for 30 minutes. Due to the weak self-rescue ability and limited expression ability of children, such accidents often find hysteresis and serious consequences. Thus, child carry-over detection is becoming a mandatory or recommended standard configuration in the automotive safety field. The prior art generally uses an in-vehicle camera to perform visual recognition to detect a legacy child, that is, by acquiring an in-vehicle image or video, whether a picture contains child features is analyzed by using a computer vision algorithm. However, the perceived efficiency of a camera is highly dependent on the ambient light, clear view, and the relative pose and position of the occupant at which it is operating. In a practical complex car scene, for example, light is dim, strong light is back light, a lens is partially blocked by clothes or articles, and when a child is in a non-standard posture such as a dead zone behind a seat or lying, the quality of visual information can be seriously reduced or even be invalid, so that the success rate of algorithm identification is greatly reduced, and the overall reliability of child identification is difficult to guarantee. Disclosure of Invention In order to solve the problems, the invention provides a method, a system, equipment and a medium for detecting children left in a vehicle. The first aspect of the invention discloses a method for detecting a child left in a vehicle, which comprises the following steps: Acquiring, for each seat, a prediction of an occupant condition of the seat by a different type of sensor; determining an initial confidence of the occupant state prediction based on the preset confidence of each sensor and the occupant state prediction; correcting the initial confidence coefficient based on a preset correction function of each sensor to obtain a correction confidence coefficient; Fusing correction confidence degrees corresponding to all types of sensors for each seat to obtain fusion confidence degrees corresponding to the seats; And judging whether children remain in the vehicle according to the fusion confidence degree corresponding to each seat. Further, the occupant condition prediction includes whether an occupant is present in the seat; and the obtaining mode of the preset confidence coefficient of the sensor comprises the following steps: And performing a supervision test on the sensor, wherein the test times exceed the preset minimum test times, the test content is whether the seat has passengers or not, and the prediction accuracy of the sensor on the passengers and the absence of the passengers is counted and used as the preset confidence. Further, the occupant state prediction includes an age bracket of the occupant; and the obtaining mode of the preset confidence coefficient of the sensor comprises the following steps: And performing supervised tests on the sensor, wherein the test times exceed the preset minimum test times, the test contents are preset age groups to which the predicted passengers belong, and the prediction accuracy of the sensor for each preset age group is counted and used as the preset confidence coefficient corresponding to each age group. Further, the obtaining manner of the correction function includes: And collecting historical detection data and corresponding real states of the sensor in a real vehicle environment, establishing a relation model between the initial confidence coefficient and the real states through a data fitting method, and determining the relation model as the preset correction function. Further, for each seat, fusing correction confidence degrees corresponding to all types of sensors, and obtaining the fusion confidence degrees corresponding to the seats includes: calculating the fusion confidence according to the following f