Endpoint Detection Accuracy in Point Cluster Reference Specification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques face challenges in accurately specifying a reference point of an object due to inaccurate end point detection, especially when the end point is not behind another object but due to unfavorable material, surface shape, or positional relations with the sensor.
Innovation Solution
An information processing apparatus comprising a detector and specifier, which detects end points and specifies a reference point if a more distant object is present within a neighboring region, improving accuracy by excluding false end points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If end point detection is performed using conventional methods, then the detection process is simple, but the measurement precision of the reference point deteriorates due to inaccurate end point detection
Solution Approach 1:
The patent applies preliminary action by performing endpoint detection on both sides of the object before final reference point determination. The detector detects endpoints in multiple directions, and the specifier selectively chooses valid endpoints based on presence information of other objects, ensuring accurate reference point specification before position and speed calculations are performed.
Solution Approach 2:
The patent introduces an intermediary mechanism through the specifier component that mediates between the raw endpoint detection results and the final reference point selection. The specifier uses presence information about other objects as an intermediary criterion to validate whether detected endpoints are true endpoints or false detections caused by occluding objects.
2Reliability
If endpoint detection is performed without considering neighboring objects, then the detection process is fast, but the reliability of reference point specification deteriorates due to false endpoint detection
Solution Approach 1:
The patent performs preliminary detection of other objects and their presence information in the vicinity of the target object before determining reference points. This preliminary action allows the system to quickly exclude false endpoints caused by occluding objects, improving reliability without requiring exhaustive re-detection.
Solution Approach 2:
The patent applies local quality by focusing detection and validation efforts specifically on regions where other objects may be present. The specifier checks presence information locally around detected endpoints, applying validation only where necessary rather than uniformly across the entire detection field, thus balancing reliability with processing efficiency.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
According to an arrangement, an information processing apparatus includes a detector(20E) and a specifier(20G). The detector(20E) is configured to detect an end point in a point cluster that constitutes a first object. The specifier(20G) is configured to specify the end point as a reference point of the first object when a second object that is more distant than the end point from the apparatus is present within a neighboring region of the end point.