Team Sahara on how to detect and monitor illegal sand mining activities

The following article covers the work of team Sahara, the winners of the “Tracking Illegal Sand Networks” challenge part of the third Mothership Mission titled The Big Blue. For more info on the program visit the website.

Sand is one of the most valuable resources we have available today. It forms the very basis of modern civilization. Over the last few years, we have seen an exponential increase in the demand for this valuable yet exhaustible resource. These alarming rates are a result of an ever-increasing building construction project and other infrastructural development. This unprecedented demand has resulted in an imbalance in the availability and accessibility of sand. Consequently, the process of acquiring it has become expensive and laborious for builders. Ergo, numerous networks have cropped across the world who conduct illegal sand mining activities. Even though we have vast expanses of sand in the Sahara and other deserts, this type of sand is not suitable for construction as the grains are rounded and cannot form an optimum bounding which is necessary for internal stability. Therefore, the sand found near rivers, lakes, and sea/ocean shorelines is more desirable because they are mostly angular. These regions are regarded as a critical resource in given its recreational, environmental, and economic importance. Yet, these are the worst affected areas due to illegal sand mining. Figure 1 shows the remote sensing images generated from Landsat data for the Poyang lake. The figures are 18 years apart and illustrate the effects of illegal sand mining activities in one of the worst affected areas in the world.

Figure 1: NASA Earth observatory image of Poyang Lake 18 years apart using the Landsat data.

The scale at which sand mining takes place highly varies from large coordinated networks dredging sand with massive machines to independent people stealing sand with shovels. This has catastrophic effects on inhabitants’ livelihood, undermines local infrastructure, and disrupts ecosystems. It is, therefore, critical for governmental & environmental agencies to observe and quantify these changes around these regions. At present, the detection and monitoring of illegal sand mining are highly tedious, in-efficient and unsafe in some cases. The open data on illegal sand mining today comes from in-situ measurements. However, this is limited to only a few spots in the world. The primary source of information on such activities comes from inhabitants and activists who would like to raise awareness. However, in most cases, their claims are disregarded due to lack of evidence. Furthermore, this large manual effort restricts the ability of governments and environmental agencies to identify and curb these illegal activities. The observations around the potentially affected areas are vital in understanding previous trends and calibrate prediction models capable of forecasting changes. Publicly available remote sensing data from Earth Observation Satellite constellations such as Landsat & Copernicus, provide a low-cost solution in obtaining long-term observations of these regions.

As part of the Mothership Missions challenge, we — Team Sahara, developed an autonomous method that combines satellite remote sensing observation with information from inhabitants & activists to rapidly detect and monitor illegal sand mining activities. Figure 2 shows the end-to-end pipeline of our proposal to tackle the above-mentioned problem.

Figure 2: An end to end pipeline for detection and monitoring of illegal sand mining networks.

The first step of the pipeline is the development of a mobile application — Sahara1 in order to provide inhabitants & activists a safe platform to raise concerns regarding illegal sand mining activities. This application will facilitate the users to request an in-depth analysis of a potential illegal sand mining area. In the next step, the available data for the “flagged” region is downloaded and processed for further analysis. We utilize Google Earth Engine to retrieve satellite images from Landsat and Copernicus constellations. Over the course of the challenge, we have built the SandSat toolbox which processes this input to facilitate detailed time series analysis of the shoreline changes to identify potential sand mining activities. Furthermore, we utilize the past data to predict the shoreline dynamics within a level of uncertainty. These results are then sent to all stakeholders (such as government agencies, media outlets, environmental agencies) including the original requester while concealing the identity of the user upon request.

We performed preliminary analysis on a specific region of Poyang Lake (Figure 1) as a use-case for SandSat. The tool was able to autonomously detect shorelines in the study area after identifying an optimal set of input parameters with the Monte Carlo Simulations. Figure 3 shows the extracted shoreline for the use case. Furthermore, with time series decomposition we were able to identify spots over

Figure 3: Output SandSat Toolbox for Poyang Lake research area. Image is taken from Sentinel-2 in June 2019. With the configurations of this model run, the dredgers were not detected as shoreline, and the full shoreline was mapped. The parameters used were: minBeachArea 16000, bufferSize 300, minLengthSl 1500 and maximumDistanceReference 400

the timeline that indicates potential sand mining activities. The time series decomposition for the shoreline per unit time is presented in Figure 4. The rate of change of shoreline is denoted as observation in the figure. With the help of this method, the observation is broken down in the following components — Trend, Seasonal & Residual. For our purposes, the trend is the most important component, since erratic changes in trend denote potential sand mining activities. This can be noticed in the figure with steep slopes on the trend. Adding, additional information such as dredging contracts, weather data, etc. into this graph can spot potentially illegal sand mining sites. Furthermore, as Landsat and Copernicus together provide coverage of over 30 years we were able to forecast the trend of the changes in shoreline for the next few years with a certain degree of uncertainty.

Figure 4: Decomposed signal.

Our business model relies on blending elements of social entrepreneurship and the lean start-up model. Over the next months, the SandSat toolbox will be developed to include three robust models: the shoreline model (from the current prototype), the bathymetry model, and the shipping tracking and detection algorithm. These models will be the backbone of a number of commercial services: retrospective sand mining information, real-time sand mining detection, automatic reporting, and contract compliance monitoring. Each service is a commercial product that will be targeted at specific customers (commercial partners), like governmental departments of conservation, natural resource authorities, and trans-national governmental organizations like the United Nations (UN). These elements comprise the commercial branch of the company. The social branch of the company consists of the User End Platform at which we communicate with the societal partners. The Sahara Mobile App is targeted for activists who want to tackle illegal sand mining, as well as concerned citizens affected by sand mining near their residential areas.

Written By: Amin Askarinejad, Sabyasachi Biswas, Somayeh Ahmadi, Srikara Datta, Sven-Arne Quist

Team Sahara won the prize for their challenge co-organized with SandStories and UCLouvain.

The Mothership is an open innovation program helping teams to develop a proof of concept and business model for solutions related to the Sustainable Development Goals. The program is co-organized by AI Lab One, Space4Good, and WorldStartup.