Vehicle perception and danger early warning method and system based on environmental sound analysis

Abstract

The invention discloses a vehicle perception and danger early warning method and system based on environmental sound analysis, and the system is characterized by comprising a sound collection module, a sound analysis module and a danger early warning module. Sound sensors arranged on the periphery of the vehicle body collect sound information around the vehicle, the sound information is transmitted to the sound analysis module, short-time energy characteristics of the sound information are extracted and compared with MFCC (Mel Frequency Cepstrum Coefficient) characteristics for analysis, and target object attributes (motor vehicles, non-motor vehicles, pedestrians and animals) are judged; the position and the relative motion relation of the sound source relative to the vehicle are judged by analyzing the time difference (TDOA) of sound reaching different position sensors, and the attributes of surrounding traffic composers and the relative position information of the vehicle and the surrounding traffic composers are displayed by a vehicle-mounted display. Whether danger occurs or not is judged according to the relative movement relation between the sensed attribute, position, movement and other information of the surrounding traffic composer and the vehicle, danger early warning is carried out on the driver of the vehicle and other traffic composers under the dangerous condition, and fusion with other types of sensor sensing information can also be carried out so that the environment sensing precision is improved.

Description

technical field [0001] The invention relates to the technical field of sound detection, in particular to a vehicle perception and danger warning method and system based on environmental sound analysis. Background technique [0002] With the continuous increase of car ownership, the traffic scene becomes more and more complex. Whether it is a manned vehicle or an unmanned vehicle, it is difficult to detect all the objects around the vehicle (including motor vehicles, non-motor vehicles, pedestrians, etc.) etc.) for observation and recognition, especially in scenes such as blind spots of vehicles, foggy days, rainy days, and nights, it is even more difficult to observe the surrounding traffic composition in detail. Based on this, a method and system for vehicle perception and danger warning based on environmental sound analysis is proposed. The traffic composition information is intuitively expressed as an image through the on-board display, and the driver is informed, and the...

AI Technical Summary

Problems solved by technology

[0004] However, the above-mentioned research has the following defects: it only uses the sound of the whistle to identify the vehicle, and filters the sounds of non-motor vehicles and pedestrians as noise, without considering factors such as non-motor vehicles and pedestrians; it is difficult to recognize only the sound of the whistle. Solve the situation when the surrounding vehicles are speeding due to faults and it is too late to honk

With the continuous increase of car ownership, the traffic scene is becoming more and more complex. Whether it is a manned vehicle or an unmanned vehicle, it is difficult to detect all the objects around the vehicle (including motor vehicles, non-motor vehicles, pedestrians, etc.) ) for observation and recognition, especially in scenes such as blind spots of vehicles, foggy days, rainy days, and nights, it is even more difficult to observe the surrounding traffic composition in detail

Images