Vehicle Object Detection Parallel Area Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional object detection devices for vehicles extend detection processing time due to a fixed detection window size and the need to resize images in multiple steps.

Innovation Solution

The device sets multiple detection areas of varying sizes around a vehicle, allowing simultaneous detection processing in parallel using pattern matching or feature analysis, with an integration unit to combine results from these areas.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a fixed-size detection window is used and the image is resized in multiple steps, then the detection processing covers the entire image, but the detection processing time is extended

Engineering Contradiction:
Improvedetection coverageVSAvoiddetection processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The image is divided into multiple detection areas with different sizes (first detection area, second detection area, third detection area) corresponding to different distance ranges from the vehicle. Each detection area is processed independently and simultaneously, eliminating the need for sequential resizing operations while maintaining comprehensive detection coverage across all distance ranges.

Inventive Principle:
Principle #1Segmentation

2Ease of operation

If the image is resized in multiple steps to adapt to fixed detection window, then the detection can be performed uniformly, but the processing complexity and time increase

Engineering Contradiction:
Improvedetection uniformityVSAvoidprocessing complexity
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

Different detection areas are assigned different sizes and processing parameters according to their specific requirements. The first detection area (near field) uses one size, the second detection area (mid field) uses another size, and the third detection area (far field) uses a third size. This allows each region to be processed optimally without requiring uniform resizing of the entire image, reducing processing complexity while maintaining detection uniformity through localized optimization.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2863361B1Object detection device for area around vehicle
Publication Date: 2020.08.12 ICHIKOH IND LTD
  • EP2863361B1 patent drawingFigure 1
  • EP2863361B1 patent drawingFigure 2
  • EP2863361B1 patent drawingFigure 3

AI summary

This object detection device for the area around a vehicle is provided with an imaging device and an image processing device. The image processing device is configured from a detection area setting unit, three area processing units, and an integration processing unit. The detection area setting unit sets a trio of small, medium, and large detection areas within a one-frame image captured by the imaging device. Detection processing for the presence of objects within the trio of small, medium, and large detection areas is carried out for each detection area simultaneously and in parallel by the three area processing units using pattern matching. The integration processing unit integrates the plurality of detection results resulting from detection processing by the plurality of area processing units and outputs the result. As a result, the present invention makes it possible to reduce the amount of time required for detection processing.