Stereo Camera 3D Object Counting via Disparity Map Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for identifying three-dimensional objects using stereo cameras often misjudge aggregated objects as single entities, leading to errors in counting, particularly when multiple objects overlap.
Innovation Solution
The method employs a Gaussian filter to enhance the profile of three-dimensional objects and the watershed segmentation algorithm to separate overlapping objects, utilizing disparity maps and calibration parameters to accurately distinguish and count individual objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional object identification methods are used, then the system is simple to operate, but the counting accuracy deteriorates when multiple objects are aggregated
Solution Approach 1:
The patent applies segmentation by dividing the identification process into distinct stages: initial object detection, profile extraction, Gaussian filtering to enhance boundaries, and watershed segmentation to separate aggregated objects. This multi-stage segmentation approach enables accurate separation of overlapping objects while maintaining systematic control over the complexity of each individual processing step.
Solution Approach 2:
The patent introduces intermediate processing steps including Gaussian filtering as a mediator between object detection and final segmentation. The Gaussian filter serves as an intermediary that enhances object profiles and boundaries before the watershed algorithm is applied, facilitating more accurate separation of aggregated objects without directly confronting the complexity of the segmentation problem.
2Measurement precision
If stereo camera systems are used to capture three-dimensional objects, then spatial position information is improved, but object separation accuracy deteriorates when objects overlap in the field of view
Solution Approach 1:
The patent leverages the three-dimensional capability of stereo cameras by extracting height information from disparity maps and projecting objects onto a plan view. This dimensional transformation allows the system to utilize vertical height data to distinguish and separate objects that may overlap in the two-dimensional image plane, thereby maintaining both spatial accuracy and separation reliability.
Solution Approach 2:
The patent applies local quality enhancement through Gaussian filtering, which selectively enhances the boundaries and profiles of individual objects based on their local characteristics. By adjusting the filtering parameters locally for each detected object, the system maintains high spatial accuracy while improving the reliability of separation for objects with varying sizes, shapes, and degrees of overlap.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for intensifying identification of three-dimensional objects identification includes utilizing a left eye camera and a right eye camera to capture a left eye image and a right eye image, calibrating the left eye image and the right eye image to generate a calibrated left eye image and a calibrated right eye image, using the calibrated left eye image and the calibrated right eye image to generate a disparity map, differentiating a three-dimensional object from a background image according to the disparity map, projecting the three-dimensional object onto a plan view, filtering noise out of the plan view to generate a filtered three-dimensional object, determining whether the filtered three-dimensional object contains at least two three-dimensional objects, and separating the at least two three-dimensional objects if the filtered three-dimensional object contains at least two three-dimensional objects.