Monocular Camera Distance Determination Using Geometric Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing distance measurement methods require expensive and high-resolution 3D sensors, which are not suitable for tasks like object classification and pose estimation, and often rely on assumptions that can lead to errors.
Innovation Solution
A system using a monocular camera as an optical sensor, combined with an electronic processing unit that receives images and predetermined identification features and geometric data of objects, to determine distances by identifying key positions in images and relating them to known object data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D sensors (TOF cameras or stereo cameras) are used for distance measurement, then distance measurement capability is achieved, but cost increases and lateral resolution decreases
Solution Approach 1:
The patent uses a monocular camera to capture a two-dimensional image that serves as a copy of the three-dimensional scene. By detecting identification features and key positions in this 2D image copy, and combining with predetermined geometric data, the system reconstructs distance information without requiring expensive 3D sensors. This copying approach allows distance measurement while maintaining low cost and high lateral resolution.
2Device complexity
If assumption-based methods are used for distance estimation with monocular cameras, then cost is reduced, but measurement accuracy deteriorates due to incorrect assumptions
Solution Approach 1:
The patent applies preliminary action by detecting and verifying identification features (such as facial features, body proportions, or object-specific markers) before performing distance calculation. Instead of directly applying assumptions, the system first identifies characteristic features of the target object/person, then uses predetermined geometric data associated with these features to calculate distance. This preliminary feature detection ensures that distance calculations are based on actual observed characteristics rather than incorrect assumptions.
3Device complexity
If monocular cameras are used instead of 3D sensors, then cost decreases and lateral resolution increases, but distance measurement capability is lost
Solution Approach 1:
The patent introduces predetermined geometric data as an intermediary between the 2D image capture and 3D distance calculation. The monocular camera captures 2D images with high lateral resolution, then the system uses predetermined geometric data (such as standard human body proportions, object dimensions, or feature spacing) as a mediator to translate 2D image measurements into 3D distance information. This intermediary approach enables distance measurement capability while maintaining the advantages of monocular cameras.
4Measurement precision
If identification features and geometric data are used for distance determination, then distance accuracy improves, but system complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-storing geometric data associated with various identification features (such as standard distances between facial features, body proportions, or object-specific dimensional relationships). During operation, the system only needs to detect the identification feature in the image and retrieve the corresponding predetermined geometric data, rather than performing complex real-time 3D reconstruction. This preliminary preparation of geometric data significantly reduces processing complexity while maintaining high distance determination accuracy.
Data Source
Figure 1
Figure 2
AI summary
A system for determining the distance of an object with respect to a reference position comprises an optical sensor located at the reference position and having a predefined field of view, and an electronic processing unit. The electronic processing unit is configured to receive an image captured by the optical sensor, receive predetermined identification features and predetermined geometric data of the object, detect the object within the optical sensor's field of view using the predetermined identification features, identify at least two key positions of the object in the image, determine the distance between the key positions in the image, and determine the object's distance with respect to the reference position using the distance between the key positions in the image and the predetermined geometric data of the object.