Dynamic Component-Wise Biomass Saturation Prediction With LiDAR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting forest biomass saturation face challenges in capturing dynamic changes over time, adapting to diverse forest types, and accounting for natural disturbances, with a reliance on ground data and limited scalability.
Innovation Solution
A method and system using non-invasive sensing techniques, integrating non-spatial knowledge from literature with remote sensing and weather data, employing machine learning models and non-fungible tokens (NFTs) to predict component-wise biomass saturation dynamically, eliminating the need for continuous ground data collection and enabling decentralized asset management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If allometric equations are used for biomass prediction, then simplicity and ease of application are improved, but generalization across diverse forest types and dynamic change capture are worsened
Solution Approach 1:
The patent transitions from static allometric equations to a dynamic deep learning model that processes temporal sequences of satellite imagery. The LSTM network captures dynamic changes in biomass over time, allowing the system to adapt to diverse forest types and growth patterns while maintaining ease of application through automated remote sensing analysis.
Solution Approach 2:
The system changes the input parameters from simple manual measurements (diameter, height) to multi-temporal satellite imagery with multiple spectral bands. This parameter transformation enables the model to capture complex forest dynamics and adapt to various forest types while keeping the user interface simple through automated analysis.
2Reliability
If process-based forest growth models are used, then mechanistic understanding and scenario analysis are improved, but device complexity and dependency on ground reference data are worsened
Solution Approach 1:
The patent replaces complex mechanical/ecological modeling systems with a data-driven deep learning approach. Instead of simulating ecological processes through complex differential equations, the system uses convolutional neural networks to directly learn biomass patterns from satellite imagery, reducing complexity while maintaining predictive accuracy.
Solution Approach 2:
The system creates a digital twin of the forest ecosystem using satellite imagery and deep learning models. This virtual representation captures forest biomass dynamics without requiring physical ground measurements or complex ecological simulations, reducing dependency on ground reference data while maintaining reliability.
3Loss of information
If component-wise biomass prediction is implemented, then detailed forest management information is improved, but data acquisition complexity and model training difficulty are worsened
Solution Approach 1:
The patent applies segmentation to divide the forest canopy into distinct components (trees, branches, leaves, understory) using deep learning. This segmentation enables component-wise biomass prediction by analyzing spectral characteristics of each layer separately, providing detailed management information while avoiding the complexity of manual ground measurement for each component.
Solution Approach 2:
The system uses satellite imagery as an intermediary to indirectly measure component-wise biomass. Instead of directly measuring each forest component on the ground, the multi-spectral satellite data serves as a mediator that captures structural and compositional information, which the deep learning model then translates into component-specific biomass estimates.
4Productivity
If dynamic biomass saturation prediction is implemented, then adaptive forest management capability is improved, but model adaptability to disturbances and spatial variability is worsened
Solution Approach 1:
The patent implements feedback mechanisms by continuously monitoring temporal changes in satellite imagery and updating biomass predictions dynamically. The system detects deviations from expected growth patterns caused by disturbances and adjusts predictions accordingly, enabling adaptive management while maintaining versatility through data-driven flexibility.
Solution Approach 2:
The system uses dynamic temporal modeling with LSTM networks to capture changing forest conditions over time. This dynamic approach automatically adapts to various disturbance scenarios (fires, pests, storms) by learning from historical patterns in the training data, maintaining versatility without requiring separate models for each disturbance type.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
The embodiments of the present disclosure herein address unresolved problems of reliance on ground reference data for machine learning (ML) based modeling for biomass allocation or process-based modeling which limits the applicability of models. Also, most of the existing biomass allocation models are static in nature. Embodiments herein provide a method and system for a dynamic prediction of component-wise biomass saturation using non-invasive sensing. Based on hierarchical integration of non-spatial, unstructured knowledge from the literature along with remote sensing and long and short-term weather (spatial), the system provides automatic spatial tree selection for biomass allocation prediction. Further, the system makes use of spatial tree locations, LiDAR data and high-resolution satellite data for component level segmentation. Further, the system predicts component-wise biomass saturation points using the temporal biomass profile and long-range weather forecast. Finally, the system provides a decentralized tree and forest asset management using NFTs and predicted biomass saturation points.