AI Sound Recognition Camera for Blind-Spot Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional camera systems for vehicles are limited in detecting objects outside their field of view, leading to potential accidents when objects approach from blind spots, such as alleys, where they are not visible to the driver or camera.
Innovation Solution
An AI-based sound recognition module and camera system that uses noise removal, voice recognition, and sound processing through a neuron artificial neural network to identify approaching objects like motorcycles, cars, or animals by their sounds, and alerts the driver, comprising multiple microphones for directionality, coherence calculation, and beamforming to filter noise, and a CMOS connector and FPGA for pattern recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If camera sensors are mounted for front, side, and rear monitoring to detect objects in the field of view, then the driving safety is improved, but objects approaching from blind spots (alleys, areas not visible to camera) cannot be detected
Solution Approach 1:
The patent transitions from two-dimensional visual detection (camera field of view) to three-dimensional spatial detection by incorporating acoustic wave propagation. Sound waves can reach areas blocked from camera view, enabling detection of objects in blind spots through the auditory dimension, thus expanding detection coverage without adding more camera mounting positions.
Solution Approach 2:
The patent replaces part of the optical detection system (camera) with an acoustic detection system (microphone array). While cameras provide visual information limited by line of sight, microphones detect sound waves that propagate around obstacles, substituting mechanical/optical limitations with acoustic capabilities to detect objects in previously invisible areas.
2Measurement precision
If multiple camera sensors are installed to cover all monitoring areas, then the detection capability is improved, but the device complexity and cost increase
Solution Approach 1:
The patent makes the microphone array system multi-functional by enabling it to perform both sound source detection and direction identification. A single acoustic sensor system achieves what would traditionally require separate visual and auditory systems, reducing overall device complexity while maintaining or enhancing detection capability.
Solution Approach 2:
The patent introduces signal processing algorithms (coherence calculation, beamforming) as intermediaries that transform raw microphone signals into meaningful detection information. These software-based intermediaries replace the need for additional physical sensors, achieving enhanced detection capability through processing rather than through hardware multiplication.
3Adaptability or versatility
If sound recognition is used to detect objects in blind spots, then the detection coverage is improved, but noise interference may cause false detection
Solution Approach 1:
The patent segments the sound detection process into distinct functional stages: noise removal, coherence calculation, and beamforming. By dividing the complex task of sound recognition into sequential processing steps, each stage can be optimized to handle specific aspects of noise filtering and signal enhancement, improving overall detection accuracy in noisy environments.
Solution Approach 2:
The patent implements feedback mechanisms through iterative signal processing where the output of each processing stage (noise removal, coherence calculation) feeds into the next stage (beamforming). This cascading feedback approach allows continuous refinement of the detection signal, with each stage correcting and enhancing the previous stage's output to minimize false detections.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Effectively recognizes and alerts the driver to approaching objects not in the camera's view, enhancing safety by preventing accidents through sound-based detection and warning systems.
Implementation Method 1
a beamforming performance unit for performing beamforming on an input signal by using the spatial filter coefficient to output a noise-processed signal
Implementation Method 2
a coherence function generation unit for calculating coherences of the input sound according to microphone intervals, respectively, calculating averages of the coherences for each identical distance
Data Source
AI summary
The AI-based sound recognition module includes: a noise removal unit for removing a noise waveform from a sound based on direction information, and outputting a result of the removal; a voice recognition unit for recognizing only a sound from a waveform output from the noise removal unit, and outputting a recognized audio signal; and a sound recognition unit for processing the audio signal to output a voice detection signal, wherein the sound recognition unit extracts a feature from an input audio signal to convert the extracted feature into a pattern vector for teaching or recognition, stores the pattern vector obtained through the conversion in a neuron library, recognizes a pattern of the pattern vector with a sound recognition model generated through library teaching, standardizes the recognized pattern, makes a global decision on the pattern, and outputs a result of the global decision as the voice detection signal.