A system and method for automatically identifying vehicle accidents
A recognition method and automatic recognition technology, which is applied in the field of automatic recognition of vehicle accident systems, achieves the effects of low cost, good scalability, and simple technical realization
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0040] Such as figure 1 A recognition method for automatically recognizing vehicle accidents shown, the recognition method includes the following steps:
[0041] S1: The control unit reads the vehicle data acquisition unit to collect vehicle data information such as brake throttle size, gear position, steering angle of the steering wheel, current vehicle speed, and vehicle light status;
[0042] S2: the control unit calculates the current running state of the vehicle based on the vehicle data information;
[0043] S3: the control unit reads the attitude estimation data collected by the measurement unit;
[0044] S4: The control unit judges whether an accident occurs through deep learning based on the current operating state data, attitude estimation data and pressure sensor data of the vehicle. If the accident does not occur, return to continue step S1, and if the accident occurs, enter step S5;
[0045] S5: Pack the running state data of the vehicle when the accident occurr...
Embodiment 2
[0052] Based on the recognition method for automatically identifying vehicle accidents in Embodiment 1, the vehicle operating state data uploaded in step S5 is combined with the actual accident of the vehicle to obtain the vehicle accident state and operating state model, and the vehicle accident state and operating state model have a significant impact on vehicle accidents. The judgment model is corrected and compensated, and the specific steps of correcting and compensating the vehicle accident judgment model are as follows:
[0053] A. Design the neural network structure and set the network structure parameters, including input layer-convolution layer-pooling layer-convolution layer-pooling layer-full connection layer-a network structure of 6 layers in total;
[0054] B. Input the two-dimensional spatial spectrogram into the convolutional neural network after setting the parameters to obtain the predicted category label, and compare it with the real category label of the sam...
Embodiment 3
[0058] A sort of figure 2 A recognition system for automatically identifying vehicle accidents is shown, the system includes a control unit, a pressure sensor, an inertial measurement unit, a vehicle data acquisition unit, and a data gateway; the vehicle data acquisition unit collects the size of the brake accelerator, gear position , the steering angle of the steering wheel, the current speed of the vehicle, the status of the vehicle lights and other vehicle data information; the inertial measurement unit is used to collect vehicle inertia data; Pressure sensors are installed on the bumper and a circle on the side of the car, surrounding the body. The pressure sensor is a sensor that converts the pressure into an electrical signal. When the position is changed by the pressure, the electrical signal of the corresponding part will change, so as to calculate the pressure value information; the control unit reads through the CAN bus system The current acceleration variation of ...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


