Artificial intelligence for identification of pollution punctual sources in arroyo Moreno, Veracruz, Mexico

The mangrove swamp of the Arroyo Moreno State Natural Reserve is of great ecological importance due to its function as a protection and containment barrier against the effect of storms and hurricanes. It is a refuge for wild flora and fauna. The basin of the Arroyo Moreno river crosses said nature reserve and the Puente Moreno housing development. There is a pollution problem in the Arroyo Moreno river basin due to the discharge of wastewater without prior treatment, mostly from the subdivision in question. This, alters the quality of the tributary. The uneven terrain of the stream bank makes it difficult to travel and represents an inconvenience to visually identify the Pollution Punctual Sources (PPS). Unmanned Aerial Vehicles (UAVs) were used to take aerial photographs of the river basin. Artificial Intelligence (AI) was trained on 160 images using the IBM® Watson Studio Visual Recognition platform. After training, the UAV was flown to detect PPS. A total of 821 photographs of the Arroyo Moreno basin were captured. Fifteen PPSs were identified in subsequent flights across the river basin with similarity percentages of up to 92%. The PPSs were georeferenced and represented using the ArGis 10.3. A map with the PPS detected by AI was obtained. The proposed methodology will serve as a basis for the development of other applications in the environmental area. This information is a support to river recovery. The aim of this research was to identify the PPSs in the Arroyo Moreno river using AI.