Vanishing Point Detection Using Monotonic Depth Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for detecting vanishing points in images face challenges such as inaccurate calculations due to unnecessary line segments and increased processing time, especially when dealing with images that contain radial subjects or signals aside from edge signals.
Innovation Solution
An image processing apparatus and method that detects vanishing points by selecting line segments with a monotonically changing distance trend in the depth direction, using distance information to improve accuracy and efficiency in vanishing point detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all detected line segments are used for vanishing point detection, then more line segments are available for calculation, but unnecessary line segments (e.g., from radial subjects or non-depth-direction segments) cause inaccurate vanishing point coordinates
Solution Approach 1:
The patent extracts and removes unnecessary line segments from the set of detected line segments before performing vanishing point detection. Specifically, it eliminates segments that do not extend in the depth direction and segments contained in radial subjects by checking distance monotonicity and angular relationships, thereby preventing these harmful segments from affecting the accuracy of vanishing point coordinate calculation.
2Stability of the object's composition
If edge signals are connected during enlargement to suppress breaks in lines, then line continuity is improved, but signals aside from edge signals are also connected, increasing processing time and reducing detection accuracy
Solution Approach 1:
The patent extracts only the necessary edge signals that constitute line segments and connects only these signals during enlargement, rather than connecting all signals. This selective approach maintains line continuity while avoiding the connection of unrelated signals, thereby reducing processing time and maintaining detection accuracy.
3Ease of manufacture
If conventional line segment selection methods are used, then processing is simpler, but unnecessary line segments remain, causing inaccurate vanishing point detection
Solution Approach 1:
The patent performs preliminary action by obtaining distance information for each line segment before vanishing point detection and using this information to select only appropriate line segments. This preliminary selection based on distance monotonicity ensures that only line segments extending in the depth direction are used, improving detection accuracy while maintaining processing efficiency through automated criteria.
Data Source
AI summary
There is provided an image processing apparatus. A line segment detecting unit detects line segments from an image. An obtainment unit obtains distance information indicating a distance in a depth direction of the image. A selecting unit selects a line segment for vanishing point detection from the line segments detected by the line segment detecting unit. A vanishing point detecting unit detects a vanishing point of the image on a basis of the line segment for vanishing point detection. For each line segment detected by the line segment detecting unit, the selecting unit selects the line segment in a case where the distance indicated by the distance information is in a monotonically changing trend along the line segment.