Stereoscopic Camera Tree Diameter Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated tree marking and dendrometry techniques face challenges in determining the ground plane, detecting individual trees, and accurately measuring tree diameters in outdoor environments, especially due to shading and color uniformity, requiring expensive equipment and computational resources.
Innovation Solution
A stereoscopic camera system coupled with a machine vision system that uses digital image data to calculate tree diameters at breast height and generate maps, employing greyscale images, pixel disparities, three-dimensional reconstruction, and convolutional neural networks for object detection and edge curve determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If traditional machine vision techniques are used for automated tree marking, then automation is achieved, but the system requires expensive and sophisticated visual sensors along with large expenditures of time and computer power to determine the ground plane
Solution Approach 1:
The patent uses digital image data copies of the forest scene processed through stereoscopic vision to determine ground planes and measure tree diameters, replacing the need for expensive sophisticated visual sensors while maintaining automation. The system creates virtual representations of the forest stand that can be analyzed computationally.
Solution Approach 2:
The patent replaces complex mechanical ground plane determination methods with computational algorithms that process digital images. Instead of using sophisticated hardware to physically determine ground planes, the system uses image processing and machine learning models to calculate ground plane positions and tree measurements from camera images.
2Difficulty of detecting and measuring
If optical sensors are used to detect individual trees in shaded forest environments, then tree detection is attempted, but the uniformity of colors and shading makes distinguishing individual trees difficult and expensive
Solution Approach 1:
The patent changes the detection parameters from color-based optical sensing to a combination of stereo vision, depth information, and machine learning classification. By transforming the problem from 2D color detection to 3D spatial analysis with depth cues, the system can distinguish trees based on their spatial relationships and geometric properties rather than color uniformity.
Solution Approach 2:
The patent adds the dimension of depth information through stereoscopic vision to the traditional 2D optical sensor data. This third dimension provides additional cues for tree detection and distinction, allowing the system to differentiate trees based on their spatial positioning and three-dimensional geometry rather than relying solely on color variations in the image plane.
3Measurement precision
If high accuracy tree diameter measurements are made in outdoor environments with irregular tree surfaces, then measurement precision is improved, but the complexity of overcoming outdoor measurement difficulties increases
Solution Approach 1:
The patent creates digital copies of tree surfaces from multiple camera viewpoints and processes these images through machine learning models to determine diameter measurements. This virtual measurement approach allows for high accuracy without requiring complex physical measurement equipment, as the system analyzes pixel disparities and three-dimensional reconstructions of tree surfaces.
Solution Approach 2:
The patent introduces machine learning models as intermediary processing layers between the camera images and the final diameter measurements. These models act as mediators that interpret the complex visual data, handle irregular tree surface geometries, and produce accurate diameter measurements without requiring direct complex physical measurement contact with the trees.
4Productivity
If a complete forest mapping system is integrated into machine vision, then comprehensive forest data collection is achieved, but the system must be transportable and inexpensive
Solution Approach 1:
The patent designs a multi-functional system where a single portable camera unit performs multiple tasks: capturing images, determining ground planes, detecting trees, measuring diameters, and generating maps. This universal approach consolidates what would traditionally require multiple separate expensive devices into one integrated portable unit that efficiently collects comprehensive forest data.
Solution Approach 2:
The system uses digital image data copies processed through computational algorithms to achieve comprehensive forest mapping without requiring expensive physical measurement equipment. The virtual reconstruction and mapping are performed through software processing of camera images, enabling portability while maintaining mapping completeness.
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
Enables accurate and efficient determination of tree diameters and geographic locations, producing a map with calculated diameters and coordinates, overcoming the difficulties of outdoor tree measurement and mapping while being transportable and cost-effective.
Implementation Method 1
First and second digital images of at least one tree, and a surrounding region, are recorded with first and second cameras, respectively... a set of pixel disparities therebetween is determined. A two-dimensional disparity map is generated from the set of pixel disparities, and three-dimensional reconstruction is performed
Implementation Method 2
First and second greyscale images are generated from the first and second digital images, respectively... At least one set of first bounding box coordinates is generated about an identified tree stem of the at least one tree in the first greyscale image
Data Source
AI summary
The method of performing dendrometry and forest mapping utilizes a stereoscopic camera system, coupled with a machine vision system, to determine the diameters at breast height of selected trees in a forest, based solely on calculations performed from recorded digital image data, as well as to generate a map showing the coordinates and calculated diameters of the selected trees in the forest. A ground plane is first determined, and then bounding box coordinates are generated about selected tree stems of the recorded images. The bounding boxes are evolved to determine stem edges of the selected tree stems, and the diameter of each tree at breast height is then determined. Geographic location data is acquired for each tree, allowing a map to be generated which shows the location of each selected tree, along with a tag representative of geolocation data and corresponding diameter data.