Description

BACKGROUND

During construction projects, the presence of protected tree species, not properly detected and identified during the preliminary project phase, could lead to delays to project timelines, costly redesign works or even bring the entire construction site activities to a halt. These situations if properly managed in advance by adopting effective and efficient techniques and methods to detect protected tree or other vegetation species, that may be rare, in danger of extinction, or of particular importance for the biodiversity equilibrium of the considered area, or that are protected by regulation, can be leveraged to transform a potential issue into opportunities to add value to the project.

In compliance with local legislation, Enel Green Power carries out manual investigations to evaluate and manage this task, combining traditional and innovative technology systems, in order to detect protected tree species in an early stage of the construction project, and to identify areas where further detailed investigations are needed. Protected tree species search and survey, traditionally carried out by expert personnel, represent a considerable effort in terms of time, and they are often not compatible with project timelines, in particular when they are carried out over large areas of hundreds or thousands of hectares.

Below are some examples of rare species alongside vegetation that is of no concern.

THE CHALLENGE

Enel Green Power (EGP) is looking for new tools to improve the environmental sustainability and biodiversity preservation of its renewable installations, as photovoltaic plants and/or wind farms. In particular, EGP is looking for solutions to be applied during the construction phase of renewable power plants to detect selected tree or other vegetation species, that may be rare, in danger of extinction, or of particular importance for the local biodiversity equilibrium of the considered area, or that are protected by regulation.

Proposals should include artificial intelligence (AI) technologies that allow the detection of selected tree species using high resolution and, if necessary, multispectral satellite images. The solution will have to identify only the desired species, without the need to recognize the surrounding species. The solution must have an excellent degree of probability of the recognition of the species, in order to mitigate as much as possible the risk during the construction phase of the plant. The solution must use, as input, satellite images in different environmental scenarios such as plains, hills or deserts, taken in different areas of the world such as Latin America, South Africa, Europe, India, etc. This list is not exhaustive, and the images could address additional Countries or scenarios. For areas partially or completely uncovered, the proposed solution should be trained within 3 to 6 months. The tree species that the solution will have to recognize belong to rare species, in danger of extinction, or of particular importance for the considered area’s biodiversity preservation. The solution is intended for use as an initial screening tool prior to confirmation of the results through manual investigation by field workers.

SOLUTION REQUIREMENTS

Submissions should address the following Solution Requirements:

The proposed solutions must:

• be characterized by very high resolution (VHR) images (< 2 meter/pixel);
• be applicable at least in Latin America, South Africa, Europe and India;
• recognize rare species, species in danger of extinction, or of particular importance for the biodiversity preservation of the considered area – minimum of one species, maximum of ten species;
• recognize the desired vegetation, the area where the vegetation species is present and/or the number of trees with a good level of confidence (90%).

Moreover, the solution should:

• be based on multispectral satellite images;
• be trained for uncovered areas within 3 to 6 months;
• provide information about the tree health;
• indicate, in certain contexts, the proximity of the tree to buildings, streets or wells.

PROJECT DELIVERABLES

The submitted proposal should include the following:

• Detailed description of the proposed method and/or technique, including (but not limited to):
  • advantages and weaknesses of the proposed solution;
  • information about the accuracy in terms of type of species, number and location of selected tree species detectable and time required for the execution of surveys;
  • evidence or calculated probability of detection, if applicable;
  • information about its applicability (i.e. type of areas/Countries, weather conditions);
  • estimate of the computing power needed to implement the proposed approach.
• Data, case studies, patent and journal references or any additional material that supports the proposed solution.

The proposal should not include any personal identifying information (name, username, company, address, phone, email, personal website, resume, etc.) or any information the Solvers may consider as their Intellectual Property they do not want to share.