Vegetation Detection from Aerial Photogrammetric Multispectral Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional approaches to forest inventory based on manual field-surveying are costly and labor-intensive due to the large area spanned by forests and variation in crown characteristics, while existing remote sensing methods lack accurate detection and delineation of individual tree crowns due to spectral variance and non-uniform illumination.
Innovation Solution
A computer-implemented method and system for vegetation detection from aerial photogrammetric multispectral data using Local Maxima detection, Fuzzy C-Means classifier with Markov-Random Field spatial-contextual model, and Gradient Vector Field snake algorithm to generate accurate fractional maps and delineate individual tree crowns, maximizing both spectral and spatial information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual field-surveying is used for forest inventory, then detection accuracy of individual tree crowns is improved, but cost and labor intensity increase significantly
Solution Approach 1:
The patent replaces manual field-surveying with an automated image processing system that uses computer vision algorithms (Local Maxima detection, Fuzzy C-Means classifier, Active Contour algorithm) to automatically detect and delineate tree crowns from aerial photogrammetric multispectral data, eliminating the need for manual measurement while maintaining high detection accuracy
Solution Approach 2:
The patent creates a digital copy of the forest canopy structure through photogrammetric multispectral data and processes this digital representation using computational algorithms to extract tree crown information, replacing the need for physical manual surveying in the actual forest
2Area of stationary object
If conventional remote sensing methods are used, then coverage area is improved, but detection accuracy of individual tree crowns deteriorates due to spectral variance and non-uniform illumination
Solution Approach 1:
The patent segments the forest canopy into individual tree crown detections by using Local Maxima detection to identify apexes and then applying Fuzzy C-Means classification and Active Contour algorithms to delineate each crown boundary separately, allowing accurate individual detection while maintaining broad area coverage through aerial photogrammetry
Solution Approach 2:
The patent applies different processing techniques to different regions of the image based on local characteristics - using Local Maxima detection for apex identification, Fuzzy C-Means for spectral classification in varying illumination conditions, and Active Contour for boundary delineation, adapting the analysis to local spectral and geometric properties
3Measurement precision
If Fuzzy C-Means classifier with Markov-Random Field spatial-contextual model is used, then classification accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary processing by first detecting apexes using Local Maxima detection and generating a height model before applying the computationally intensive Fuzzy C-Means classifier with Markov-Random Field, using these preliminary results to guide and constrain the subsequent classification process
Solution Approach 2:
The patent uses the height model and apex detection results as intermediary information that guides the Fuzzy C-Means classification process, providing spatial and structural context that simplifies the classification task while maintaining high accuracy
4Manufacturing precision
If active contour algorithm is used for delineation, then boundary accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary apex detection and height model generation before applying the active contour algorithm, using these pre-computed geometric constraints to guide the boundary delineation process and reduce the iterative processing time required by the active contour method
Solution Approach 2:
The patent applies the active contour algorithm locally around detected apexes and ridge features rather than processing the entire image uniformly, focusing computational effort on specific regions where crown boundaries are most challenging to delineate
Data Source
AI summary
Systems and methods for vegetation detection from aerial photogrammetric multispectral data. The method includes: detecting apexes in a height model using Local Maxima (LM) detection; detecting vegetation as detected apexes; performing orthorectification to derive an orthomosaic; generating a fractional map of a vegetation class by applying a Fuzzy classifier on the orthomosaic using the detected vegetation to define the class; generating a binary ridge map using the height model to identify ridges; combining the binary ridge map and the fractional map of the vegetation class to generate a ridge integrated fractional map; performing delineation of individual vegetation on the ridge integrated fractional map on a vegetation class using an active contour algorithm; and outputting the delineated vegetation.