Image Processing Apparatus Road Surface Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies face challenges in accurately identifying road surfaces and detecting objects like other vehicles due to the limitations of gray-scale maps, which are susceptible to brightness variations and result in false sensing.
Innovation Solution
An image processing apparatus that includes an object detection section performing convolution computation, a feature map validation section, a time series validation section, and a detection result correction section, which together enhance the accuracy of object detection by correcting detection results based on feature maps and time series validation, enabling the apparatus to output warning signals or vehicle control signals for the own vehicle.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If gray-scale map is used for road surface identification, then processing simplicity is improved, but measurement precision deteriorates due to narrow dynamic range and brightness susceptibility
Solution Approach 1:
The patent transforms the image data from gray-scale representation to color space representation (HSV or LAB coordinates). This parameter change allows the system to utilize color information which provides a wider dynamic range and better brightness invariance, thereby improving road surface identification accuracy while maintaining processing efficiency through standardized color space transformations.
Solution Approach 2:
The patent introduces additional dimensional information by transitioning from 2D gray-scale intensity to 3D color space coordinates (H, S, V or L, A, B). This dimensional expansion provides more discriminative features for road surface detection, enabling the system to distinguish road surfaces from backgrounds more accurately by utilizing chromaticity information in addition to luminance.
2Adaptability or versatility
If background objects are included in the captured image, then comprehensive scene monitoring is improved, but object detection accuracy deteriorates due to false sensing
Solution Approach 1:
The patent applies different processing strategies to different regions of the image. Road surface regions are identified using color space thresholds specific to asphalt and concrete characteristics, while other regions are processed differently. This localized approach allows the system to maintain high detection accuracy for vehicles on road surfaces while properly handling background regions that should not trigger false alarms.
Solution Approach 2:
The patent introduces road surface identification as an intermediary step between image capture and object detection. By first segmenting the image to identify road surface regions and then performing vehicle detection only within those regions, the system effectively filters out background objects that would otherwise cause false sensing, while still maintaining comprehensive monitoring of the entire scene.
3Productivity
If conventional object detection is used without validation, then detection speed is improved, but reliability deteriorates due to false sensing from background
Solution Approach 1:
The patent performs preliminary road surface identification and region segmentation before executing the main object detection algorithm. By pre-processing the image to identify and mask non-road regions, the system eliminates potential false detection sources in advance, ensuring that the subsequent detection algorithm only processes relevant regions, thus maintaining high reliability without significant speed penalty.
Solution Approach 2:
The patent implements a feedback mechanism where detection results are validated against the previously identified road surface regions. If a detected object falls outside the road surface region or exhibits characteristics inconsistent with valid detections, the system can reject or re-evaluate the detection result. This feedback loop significantly improves detection reliability by filtering out false positives while maintaining efficient processing through targeted validation.
Data Source
AI summary
An image processing apparatus includes: an object detection section that performs convolution computation on an input image based on a captured image obtained by capturing the image with a camera, and that detects an object; a feature map validation section that performs feature map validation validating a likelihood that the input image contains the object on the basis of a feature map obtained by the convolution computation; a time series validation section that performs time series validation validating a result of the feature map validation performed by the feature map validation section in time series; and a detection result correction section that corrects a detection result about the object output by the object detection section on the basis of a result of the time series validation performed by the time series validation section.