Target Search Apparatus Region Segmentation Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current target searching methods require operators to manually identify detection targets from wide-angle camera views, involving repetitive zooming and angle adjustments, which is operationally burdensome and inefficient.
Innovation Solution
A target searching apparatus and method that extracts an object region from a displayed image, calculates image feature amounts within and outside this region, and identifies the object's presence by dividing the region into smaller areas, allowing for automatic zooming and view adjustments to confirm the detection target.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual searching is performed by operator viewing wide-angle camera pictures, then the operator can search over a wide range, but the operator burden increases and efficiency decreases due to repetitive zooming and angle adjustments
Solution Approach 1:
The patent segments the image processing task by dividing the wide-angle image into multiple regions and analyzing each region separately to identify potential detection targets. This automation of region analysis reduces the operator's manual workload while maintaining comprehensive search coverage.
Solution Approach 2:
The system performs self-service by automatically extracting object regions, calculating image feature amounts, and identifying potential detection targets without requiring continuous manual intervention. The automated processing of image features and region extraction reduces operator burden while improving detection efficiency.
2Measurement precision
If the operator zooms in to confirm possible detection targets, then identification accuracy improves, but the time required for target confirmation increases
Solution Approach 1:
The patent applies preliminary action by pre-processing the wide-angle image to extract object regions and calculate image feature amounts before the operator needs to confirm a target. This preliminary analysis identifies potential detection targets and prepares region information in advance, reducing the time required for confirmation while maintaining identification accuracy.
Solution Approach 2:
The system replaces the mechanical zooming operation with an automated image processing mechanism that calculates image feature amounts and extracts object regions. This substitution eliminates the need for manual zooming while providing accurate target identification through computational analysis of image features.
3Area of stationary object
If the operator manually adjusts camera angles to search different areas, then comprehensive coverage is achieved, but the operational complexity increases
Solution Approach 1:
The patent transitions from spatial dimension (manual camera angle adjustment) to computational dimension (image feature analysis). By processing the entire wide-angle image and dividing it into regions for feature analysis, the system achieves comprehensive coverage through image processing rather than mechanical camera movement, reducing operational complexity.
Solution Approach 2:
The system applies universality by using a single wide-angle camera view that serves multiple functions: both comprehensive area coverage and detailed target identification. The automated region extraction and image feature calculation enable the same image to be analyzed at multiple levels without requiring multiple camera adjustments or positions.
Data Source
AI summary
A target searching apparatus includes a display, an object region extractor, a feature amount calculator, a second feature amount calculator, and an object present region extractor. The object region extractor extracts, from a display image, an object region including an identification object. The feature amount calculator calculates in-region and out-region representative values of the image feature amount that respectively are representative values inside and outside the object region in the display image. The second feature amount calculator calculates a representative value of the image feature amount in each of a plurality of small regions into which the object region is divided. The object present region extractor extracts, from the plurality of small regions, one or more small region having the representative value that is closer to the in-region representative value than the out-region representative value, as an object present region in which the identification object is present.