Autonomous Vehicle Object Detection With Field-of-View Image Cropping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine vision systems for autonomous vehicles face challenges in achieving accurate and efficient object detection due to high computational requirements and the need for downsampled images, which reduce detection range and accuracy.
Innovation Solution
The method involves determining a priority field of view in an image, cropping it to high resolution, and downsampling the remaining image, allowing for efficient processing by a detector that combines the high-resolution crop with the low-resolution image to enhance object detection accuracy and range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high resolution images are used for object detection, then detection accuracy and range are improved, but computational requirements and processing time increase significantly
Solution Approach 1:
The patent divides the image processing task into two segments: a first downsampled image for detecting distant objects and a second cropped image at higher resolution for detecting closer objects. This segmentation allows the system to process different regions at appropriate resolutions, improving both detection accuracy and processing efficiency by avoiding unnecessary high-resolution processing of entire images.
2Productivity
If downsampled images are used to reduce computational requirements, then processing speed is improved, but detection range and accuracy deteriorate
Solution Approach 1:
The patent applies local quality by processing different regions of the image at different resolutions. The first downsampled image provides sufficient quality for detecting distant objects, while the second cropped image at higher resolution provides the necessary detail for detecting closer objects. This local differentiation of image quality optimizes both processing speed and detection accuracy for different spatial regions.
3Reliability
If the entire high resolution image is processed, then all objects are detected with sufficient detail, but computational resources are excessively consumed
Solution Approach 1:
The patent extracts and processes only the necessary portions of the image at high resolution. By cropping a specific region from the original image to create the second image, the system extracts only the relevant high-detail area that requires close-object detection, avoiding the computational expense of processing the entire high-resolution image while maintaining detection completeness for objects that require detailed analysis.
Data Source
AI summary
Systems and methods for enhanced object detection for autonomous vehicles based on field of view. An example method includes obtaining an image from an image sensor of one or more image sensors positioned about a vehicle. A field of view for the image is determined, with the field of view being associated with a vanishing line. A crop portion corresponding to the field of view is generated from the image, with a remaining portion of the image being downsampled. Information associated with detected objects depicted in the image is outputted based on a convolutional neural network, with detecting objects being based on performing a forward pass through the convolutional neural network of the crop portion and the remaining portion.