Autonomous Vehicle Object Detection Using FOV Cropping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In the field of machine vision for autonomous vehicles, existing object detection methods face challenges due to high computational requirements, which necessitate downsampling of images, reducing detection accuracy and range. Current detectors are either slow and accurate or fast but less accurate, failing to effectively detect objects at both close and far distances.
Innovation Solution
A method that involves cropping a high-resolution field of view from an image and downsampling the remaining portion, allowing both images to be processed together by a detector to enhance computational efficiency and detection accuracy, enabling long-range object detection while maintaining high accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high resolution images are used for object detection, then detection accuracy is improved, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent divides the image processing task into two segments: a first downsampled image for general object detection and a second cropped high-resolution image for detailed verification. This segmentation allows the system to process images at different resolutions for different purposes, balancing accuracy and speed by applying high computational resources only where necessary rather than processing the entire image at full resolution
Solution Approach 2:
The patent applies different quality levels to different parts of the image by cropping a specific region of interest from the original high-resolution image. The cropped region containing potential objects is processed at full resolution for accurate detection, while the rest of the image is handled through downsampling. This local quality approach ensures high detection accuracy is applied only to relevant areas, reducing overall computational burden
2Productivity
If images are downsampled to reduce computational requirements, then processing speed is improved, but detection range and accuracy decrease
Solution Approach 1:
The patent performs preliminary downsampling to quickly identify regions containing objects, then uses these results to guide subsequent cropping of high-resolution regions. This preliminary action at low resolution enables the system to rapidly screen the image and focus computational resources on relevant areas, achieving both speed and accuracy by preparing the processing strategy in advance
Solution Approach 2:
The downsampled image acts as an intermediary that bridges the gap between full-resolution and processed images. It provides a computational shortcut for initial object localization, whose results then inform the cropping of high-resolution regions. This intermediary representation enables the system to trade some detail for speed in the initial phase, then recover accuracy in the verification phase
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.