Start a discussion
For participation in the discussion you need to Register/Login. Start a discussion by introducing yourself, posting relevant texts, photos, links, documents, news, etc. Add proper KeyWords (#): they are crucial for future topic searches.
Guide: How to use CEETO Network platform1.44 MB PDF uploded 5 years ago
Search
Login
Chiara Rognoni 5 years ago
CEETO Referent at Ente di gestione per i Parchi e la Biodiversità Emilia Centrale
CEETO pilot actions in Emilia Centrale Parks
Lago Santo Modenese area, within the High Modenese Apennines Regional Park, and Salse di Nirano Nature Reserve are two pilot areas managed by the Emilia Centrale Parks Management Authority. Both areas are affected by a high tourist presence which, along with inadequate visitors´ behaviour, cause damage to natural values and overall experience of the protected area. Therefore, in Lago Santo Modenese area tourist flows were monitored via the number of sold parking tickets and questionnaires to define measures for better spatial and time distribution of tourists’ visits in the area. Also, sustainable traffic measures and raising knowledge and awareness about the value of the area were implemented. In Salse di Nirano area, to protect the Zone A (integral protection zone) of the Reserve, a Video Content Analysis (VCA) system was installed to detect the type of the use and threats the Reserve must face from the point of view of sustainable tourism planning. The video of the system can be seen here: https://youtu.be/6mbOOV_c0FI
#CEETOPilotAction #EnteParchiEmiliaCentrale #High Modenese Apennines #Italy #Lago Santo Modenese #monitoring #ParchiEmiliaCentrale #ProtectedArea #RegioneEmiliaRomagna #RiservaNaturale #SalseDiNirano #SustainableTourism #VCA #VideoContentAnalysisMauro Generali 5 years ago
CEETO Project PP01 Communication Manager at Regione Emilia Romagna - Servizio Aree Protette, Foreste e Sviluppo della Montagna
CEETO Project - Video-Monitoring of the Salse di Nirano Nature Reserve to monitor and prevent Zone A intrusions.
NEMOS (Nature rEserve MOnitoring System) VCA Experimental System
The Video Content Analysis (VCA) system, financed and tested through the Interreg CEETO Project, is composed of 3 cameras connected to an automatic data processing and extraction unit, using Artificial Intelligence algorithms. The system, in addition to monitoring (quantifying and qualifying) the real use of the area, also documents the main factors of threat for the Natural Reserve of Salse di Nirano, to support the design of possible protection measures.
As can be seen from the explanatory video (see link below), the system detects the transits of vehicles and people along the access road to Zone A of integral protection (zone particularly fragile by the geomorphological and phytological point of view) and acquires the largest number of data on episodes of physical barriers overcoming (by both people and animals), and consequent incorrect use and invasion of Zone A. The data, processed on site and sent remotely to the Park Authority and to the Municipality, will be used both to quantify the extent of the problem and to plan appropriate defence actions, both active (strengthening of physical barriers) and passive (through posters and environmental education).
The recorded images will also be useful for the study of the mud-volcanoes morphological evolution and will be integrated with the other ongoing monitoring by the University of Modena and Reggio Emilia.
The innovative VCA system was installed in the Cà Rossa Visitor Centre area at the beginning of July 2019 and it’s acquiring data continuously from 5.30 AM to 9.30 PM. It will continue to work for at least one year after the end of CEETO Pilot Action (which ended at the end of September 2019).
This is a demo Video: https://youtu.be/6mbOOV_c0FI.
This other one is a "real-life" acquisition result video collage and a preview of the web-managing interface for the DB collected data analysis: https://www.youtube.com/watch?v=qN5qA_Ov-Mw.
#CrossFences #EnteParchiEmiliaCentrale #FioranoModenese #Italy #monitoring #ParchiEmiliaCentrale #PilotAction #ProtectedArea #Protection #RegioneEmiliaRomagna #RER #RiservaNaturale #SalseDiNirano #TouristMonitoring #VCA #VideoContentAnalysis #VulcaniDiFango 3 commentsAna Krvarić 5 years ago
Project Officer at WWF Adria
Interesting action!
I have a few questions:
- 1) How one can distinguish from people/animals walking on the right area (like the street) and those who are crossing the fences and walking in the Zone A of the Protected Area?
- 2) What are power sources for the cameras?
- 3) Is this method accompanied with some other method, for example with interviews in order to know what visitor’s expectations in the area are?
Rudy Melli 5 years ago
CEO at Vision-e Srl
Hello Ana, I'm Rudy from Vision-e, the author of the VCA and I gladly reply to the first two questions.
We use a neural network applied to the images to detect objects (humans, vehicles, animals). On each frame the network extract the bounding box (the colored rectangles into the video) containing the object and its class (people, car, truck, bike, bird, dog, ...). In this way we know where are humans and animals every seconds. The perimeters of Zone A of the Protected Area have been defined manually for each camera, unfortunately they are not shown in the video but are close to the fences. With this data it becomes easy to calculate who and when is in the Zone A.
The cameras are IP and POE (Power Over Ethernet) so the power supply is injected in data cables coming from Cà Rossa Visitor Center.
These are a low power cameras with less that 3W for each. The brain of the system is not a PC but an embedded board low power consuming (peak < 15W). This hardware configuration is designed to easily switch to solar panels, making the system autonomous, in order to be replicated in areas where a power source is not available.
Mauro Generali 5 years ago
CEETO Project PP01 Communication Manager at Regione Emilia Romagna - Servizio Aree Protette, Foreste e Sviluppo della Montagna
This innovative sperimentation of VCA in "almost natural" environment, took some energy and time, so, at the moment, no other survey has been carried out. This method can be easily integrated with other survey methods, like questionnaire. Furthermore, its authomatic and quantitative approach, make it easy to assess the effect of every managerial activity like education/sensibilization, placing of plates, etc.
Andrea Solić 5 years ago
Programme Manager at WWF Adria
wouuu
Andrea Solić 5 years ago
Programme Manager at WWF Adria
Mauro, it would vbe great to get reply to this questions. Maybe PA practitioners and rserchers already use this method for wildlife monitoring? :-)
Mauro Generali 5 years ago
CEETO Project PP01 Communication Manager at Regione Emilia Romagna - Servizio Aree Protette, Foreste e Sviluppo della Montagna
So far the recognition algorithms are "educated" on more "urban" environments, but, if there is interest in the matter, I believe that the Neural Network (AI) can be trained to recognize also different species of animals and therefore be used also for the automatic monitoring of the passages of various species. Obviously it will not be able to go into great detail (like distinguishing deer, roe deer and fallow deer), but a roe deer from a wild boar, a badger or a wolf, I think it is certainly possible.
Rudy Melli 4 years ago
CEO at Vision-e Srl
I confirm Mauro's answer, by teaching the neural network with the appropriate examples it is possible to make them learn to distinguish numerous animal species.
Rudy Melli 4 years ago
CEO at Vision-e Srl
Here you can watch a video with details of the features of Nemos system:
https://www.youtube.com/watch?v=qN5qA_Ov-Mw
Mauro Generali 4 years ago
CEETO Project PP01 Communication Manager at Regione Emilia Romagna - Servizio Aree Protette, Foreste e Sviluppo della Montagna
Thank you Ing. Melli. I'll update the post to include the new video of "real life" acquisition.
Latest uploaded documents
Visitors Count_Guidance for protected areas on the economic analysis of visitation8.55 MB PDF uplodaded 3 years ago
7.02 MB PDF uplodaded 3 years ago
2.18 MB PDF uplodaded 3 years ago
3.97 MB PDF uplodaded 3 years ago
2.05 MB PDF uplodaded 4 years ago