Object Detection Using Density-Weighted Mask for Point Clouds
Solution Overview
Problem
Object detection and dimensioning in environments with extraneous data is slow and error-prone due to the time-consuming segmentation process of separating the target object from its surrounding environment in point cloud data.
Innovation Solution
A method and device that use a mask with decreasing density away from its central point to select a subset of depth data points, facilitating the detection of a cluster representing the target object, thereby reducing computational complexity and increasing accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a segmentation process is used to separate the target object from surrounding environment in point cloud data, then object detection accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent extracts only the relevant data points needed for object dimensioning by using a bounding box approach. Instead of segmenting the entire point cloud, the system identifies and extracts only the points that define the object's boundaries, eliminating the need for time-consuming segmentation while maintaining detection accuracy.
Solution Approach 2:
The patent segments the point cloud data by applying a bounding box that encloses the target object. This segmentation approach divides the data into relevant (points within the bounding box) and irrelevant (points outside) portions, enabling fast extraction of object dimensions without processing the entire point cloud.
2Reliability
If traditional preprocessing algorithms are applied to segment point cloud data, then object detection reliability is improved, but computational complexity increases
Solution Approach 1:
The patent eliminates complex preprocessing algorithms by directly extracting object dimensions from point cloud data using a bounding box method. This approach removes unnecessary computational steps while maintaining reliable object detection by focusing only on the essential data points needed for dimensioning.
3Measurement precision
If all point cloud data is processed to ensure complete object coverage, then measurement completeness is improved, but processing speed decreases
Solution Approach 1:
The patent extracts only the necessary data points for dimensioning by using a bounding box approach. Instead of processing all point cloud data to ensure complete coverage, the system efficiently identifies and extracts only the points that define the object's boundaries, achieving both completeness and speed.
Solution Approach 2:
The patent applies partial action by processing only the subset of data points that are relevant to object dimensioning. By using a bounding box to limit processing to only the necessary region, the system avoids excessive processing of unrelated data while ensuring complete object coverage.
Data Source
AI summary
A method for detecting and dimensioning a target object includes: obtaining depth data representing a target object; defining a mask having a structure which decreases in density away from a central point of the mask; overlaying the mask on the depth data and selecting a subset of the depth data comprising data points which contact the mask; detecting a cluster of data points from the subset; detecting, based on the cluster, the target object; and outputting a representation of the target object.