Remote sensing forest valuation product

• Grant number: 20/01468-0
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: July 01, 2020
• End date: June 30, 2022
• Field of knowledge: Agronomical Sciences - Forestry Resources and Forestry Engineering - Technology and Use of Forest Products
• Agreement: National Research Council of Canada
• Principal Investigator: Rafael Antonio da Silva Rosa
• Grantee: Rafael Antonio da Silva Rosa
• Company: Visiona Tecnologia Espacial S/A
• CNAE: Produção florestal - florestas plantadas ; Produção florestal - florestas nativas ; Atividades de apoio à produção florestal
• City: São José dos Campos
• Associated scholarship(s): 21/01120-6 - Remote sensing forest valuation product, BP.TT

Abstract

Forest companies and assets are regularly traded on international markets, and their valuation is absolutely necessary for land acquisition and prediction of. Accurate and reliable forest valuation is actively sought by investors globally and is important for the definition of public policies for the sector. We are proposing to to develop a forest market intelligence product to provide accurate and reliable valuation of forest assets. The lack of reliability of the status of commercial forests partially results from a more complex nature of forests when compared to other commodities (e.g., sugarcane); one needs to know forest location, if it is planted or native, species, age, volume, and other attributes that, together, can provide a complete depiction of forests now and in the future. The challenges we face to develop a value-add forest valuation product include: 1) satellite data acquisition must cover the entire region; 2) specifying what data and information the market requires; 3) adapting analytics for a situation where validation plots are scarcer and less standardized; 4) analyses must transition between different spatial scales, from local to global, without losing important information; 5) information must be presented readily, and straightforwardly; and (6) adapting infrastructure in both companies to enable the automated delivery of this service. We intend to explore freely available satellite imagery - e.g., Landsat-8, and Sentinel-1 and 2 (radar and optical) - as well as higher resolution imagery with desired details and either incorporate this product into a platform to offer this solution to the markets in Brazil and Latin America. A library of plot data will be used to create predictive models with universal characteristics, which, if not suitable for a high-resolution inventory, may be suitable to creating information for market intelligence that is sufficiently reliable. Developing an automated land cover classification (LCC) system is key to this project's success. We need to be able to, at least, classify the landscape in the following classes: (1) forest/non-forest; (1.1) if forest, determine if plantation or native; (1.1.1) if plantation forests, quantifying key attributes, and applying our existing analytics; (1.1.2) if native forests - determining and quantifying other key attributes; (2) Infrastructure; (2.1) Roads; (2.2) Constructions; and (3) Water bodies. Another goal was to associate the classification results to other data - such as topography, roads, water bodies, cities boundaries - from the classification and from public data sources. Three proposed activities push both companies beyond their comfort zones: 1) Perform a market study to specify an optimum viable product as desired by future clients; 2) Create an economic modeling to provide, as much as possible, objective economic information about a forested area currently and in the future; and 3) Acquiring the know-how to process, visualize and analyze large datasets (Big Data) in a timely manner using cloud computing environments. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

DALLAQUA, F. B. J. R.; ROSA, R. A. S.; SCHULTZ, B.; FARIA, L. R.; RODRIGUES, T. G.; OLIVEIRA, C. G.; KIESER, M. E. J.; MALHOTRA, V; DWYER, T.; WOLFE, D. S.; et al. FOREST PLANTATION DETECTION THROUGH DEEP SEMANTIC SEGMENTATION. XXIV ISPRS CONGRESS: IMAGING TODAY, FORESEEING TOMORROW, COMMISSION III, v. 43-B3, p. 8-pg., 2022-01-01. (20/01468-0)