Robot Cleaner Boundary Detection Using Brightness-Based Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot cleaners face challenges in accurately detecting boundaries between objects, such as the floor and walls or obstacles, which are crucial for correct movement and collision avoidance, as existing methods are inefficient in distinguishing these boundaries from images.
Innovation Solution
A method for controlling a robot cleaner that involves acquiring an image, selecting a seed pixel, calculating brightness differences between upper and lower regions within predetermined detection areas, and using these differences to detect and control movement based on the detected boundaries, aided by a light projection unit for patterned light projection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the robot cleaner analyzes all pixels in the acquired image to detect boundaries, then the detection coverage is comprehensive, but the detection time increases and efficiency decreases
Solution Approach 1:
The patent divides the image into multiple detection areas (first detection area, second detection area, third detection area) with different detection strategies. The first detection area uses comprehensive pixel analysis for high accuracy, while the second and third detection areas use simplified methods based on brightness thresholds and seed pixels, reducing overall processing time while maintaining adequate detection coverage.
Solution Approach 2:
Different detection methods are applied to different regions of the image based on their importance and characteristics. The first detection area (containing the robot and immediate surroundings) receives intensive processing with full pixel analysis, while peripheral areas use faster, threshold-based detection, optimizing the balance between accuracy and speed for each local region.
2Measurement precision
If the robot cleaner processes all detection target pixels to identify boundaries, then the boundary detection accuracy is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent applies full pixel analysis only to the first detection area where boundaries are most critical, while using partial action (threshold-based methods) for the second and third detection areas. This selective approach achieves sufficient boundary detection accuracy without the excessive computational complexity of processing all pixels with the same intensive method.
Solution Approach 2:
The patent performs preliminary processing by identifying seed pixels and calculating brightness thresholds before conducting full boundary detection. This preliminary action filters out obvious non-boundary regions, reducing the number of pixels that require intensive processing and simplifying the overall computational complexity.
3Reliability
If the robot cleaner uses a comprehensive detection method for all areas, then the boundary detection reliability is improved, but the detection speed decreases
Solution Approach 1:
The detection process is segmented into multiple areas with different reliability requirements. The first detection area uses the most reliable comprehensive pixel analysis method, while the second and third detection areas use faster, less reliable methods that are sufficient for their respective purposes, achieving overall reliable detection without sacrificing speed.
Solution Approach 2:
The patent dynamically adjusts the detection method based on the detection area and observed characteristics. When boundaries are clearly identified in the second detection area using threshold methods, the system transitions to faster detection without compromising reliability. The detection strategy adapts to the actual scene, maintaining reliability while optimizing speed.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method enhances boundary detection accuracy and reduces detection time by specifying detection regions based on obstacles, allowing the robot cleaner to effectively navigate and avoid collisions with walls or obstacles.
Implementation Method 1
acquiring an image of an area in front of a main body
Implementation Method 2
a light projection unit for downward projecting patterned light including a horizontal segment to an area in front of the main body
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
A robot cleaner and a control method thereof according to the present invention select a seed pixel in an image of an area in front of a main body, obtain a brightness difference between an upper region and a lower region obtained by dividing a predetermined detection area including each of neighboring pixels of the seed pixel and select a pixel belonging to a detection area having a largest brightness difference as a pixel constituting the boundary between objects indicated in the image. Accordingly, the boundary between objects present within a cleaning area can be detected rapidly and correctly through an image.