Parking Line Marker Pairing for Fast, Accurate Spot Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicular driver assist and autonomous driving systems face inaccuracies and inefficiencies in parking spot detection due to reliance on traditional AI algorithms for segmenting parking line markers, which are time-consuming and prone to errors.
Innovation Solution
The implementation of a deep neural network (DNN) based system that preprocesses camera images into a bird's-eye-view format, segments parking line markers, and performs binary image processing to accurately detect and pair parking spots, utilizing an encoder-decoder network for feature extraction and upsampling, and line fitting to enhance detection speed and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional AI algorithms are used for segmenting parking line markers, then the system can detect parking spots, but the detection process is time-consuming and inaccurate
Solution Approach 1:
The patent segments the parking detection task into distinct processing stages: image acquisition, preprocessing, deep neural network segmentation, binary image processing, and parking spot detection. This segmentation allows each stage to be optimized independently, with the DNN specifically handling the complex line marker segmentation to improve accuracy while maintaining efficiency.
Solution Approach 2:
The patent introduces a bird's-eye-view (BEV) image as an intermediary representation between the original camera images and the final parking spot detection. The BEV image provides a top-down perspective that simplifies the geometric relationships of parking line markers, making segmentation more accurate and efficient while serving as a bridge for the DNN processing.
2Measurement precision
If complex AI algorithms are used to improve parking spot detection accuracy, then detection precision improves, but system complexity increases
Solution Approach 1:
The algorithm is segmented into modular components: image acquisition module, preprocessing module, DNN segmentation module, binary image processing module, and parking spot detection module. This modular architecture reduces overall system complexity by allowing each module to be developed, tested, and optimized independently while maintaining clear interfaces between them.
Solution Approach 2:
The preprocessing module performs preliminary actions on the input images before they reach the DNN segmentation module. This includes geometric corrections, perspective transformations to BEV, and enhancement of line marker features, which simplifies the subsequent segmentation task and reduces the complexity of the DNN architecture needed to achieve high accuracy.
3Adaptability or versatility
If multiple camera images are processed to generate a 360-degree view, then parking spot detection coverage improves, but processing time increases
Solution Approach 1:
The patent merges multiple camera images into a single integrated bird's-eye-view representation that provides comprehensive 360-degree coverage. By combining the information from multiple cameras into one unified BEV image, the system achieves complete detection coverage while reducing the total processing time compared to analyzing each camera image separately.
Solution Approach 2:
The patent transforms 2D camera images into a 3D bird's-eye-view representation, adding a dimensional perspective that provides comprehensive spatial coverage. This dimensional transformation allows the system to capture parking spots from all angles simultaneously in a single processed view, improving coverage without proportionally increasing processing time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Parking spot detection methods and systems that detect and pair parking line markers, improving the accuracy and speed of parking spot detection in vehicular driver assist (DA) and autonomous driving (AD) applications. These methods and systems incorporate three major steps: (1) preprocessing - one or more standard camera images are input into the algorithm and a bird's-eye-view (BEV) image is output from the algorithm; (2) deep neural network (DNN) segmentation of the parking line markers - the BEV image is input into the algorithm and a binary image is output from the algorithm (with, e.g., the parking line markers areas displayed in white and the background displayed in black); and (3) binary image processing - the binary image is input into the algorithm and detected and paired parking line markers representing the parking spots are output from the algorithm.