Imaging Device Stereo Monocular Region Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing imaging devices face inefficiencies in computational load when detecting multiple objects at varying distances, as they require both monocular and stereo processing, which increases processing complexity and load.
Innovation Solution
The implementation of a multi-imaging unit system where distance calculations are performed using disparity between images in specific regions, allowing for efficient region control and reduced processing load by designating regions for distance detection and object detection separately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If monocular processing is used for all objects, then computational load is reduced, but sensing precision deteriorates for near objects
Solution Approach 1:
The imaging region is segmented into a first region (distant objects) and a second region (near objects). Different processing methods are applied to each region: monocular processing for the first region and stereo processing for the second region. This segmentation allows the system to optimize computational load while maintaining high sensing precision for near objects that require it.
2Measurement precision
If stereo processing is used for all objects, then sensing precision is enhanced, but computational load increases
Solution Approach 1:
The imaging region is divided into a first region for distant objects and a second region for near objects. Stereo processing is selectively applied only to the second region where high precision is critical, while monocular processing is used for the first region to reduce computational load. This selective segmentation resolves the contradiction between precision and computational efficiency.
Solution Approach 2:
Different processing qualities are applied to different regions of the imaging area. The second region (near objects) receives high-quality stereo processing with full computational resources, while the first region (distant objects) receives optimized monocular processing. This local quality differentiation ensures maximum precision where needed while minimizing overall computational load.
3Adaptability or versatility
If both monocular and stereo processing are performed for multiple objects, then comprehensive object detection is achieved, but processing complexity increases
Solution Approach 1:
Objects are segmented into two categories based on their position: distant objects in the first region and near objects in the second region. This segmentation enables the system to use appropriate processing methods for each category, achieving comprehensive detection without uniformly applying complex stereo processing to all objects, thus reducing overall processing complexity.
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 approach enables enhanced sensing precision and reduced computational load, even when multiple objects are present, by optimizing processing regions and sharing hardware resources.
Implementation Method 1
imaging units that image a subject and output imaging images
Data Source
AI summary
The purpose of the present invention is to provide an imaging device whereby enhanced sensing precision and reduced computational load can both be achieved at the same time even when a plurality of objects as sensing subjects are present. In order to achieve this purpose, the present invention has: a first object detection processing unit having a plurality of imaging units, the first object detection processing unit performing first object detection processing for sensing, by stereo processing, a distance for each of a plurality of image elements in partial region of interest in an image acquired by the imaging units, then extracting a group of the image elements on the basis of the sensed distances and detecting an object; and a second object detection processing unit for performing second object detection processing for detecting the distance of the object by stereo processing for a partial region of another region of interest in the image.