Monitoring of boat navigation in the dalyan channel

Abstract

Specially protected areas (SPAs), are generally attractive for visitors and tourists. Yet, several human activities can create pressure on precious ecosystems in SPAs. Especially, boat traffic is documented to have various negative effects on deltaic-lagoonal ecosystems. Hence, in order to prevent these environments from certain threads and manage their sustainability, monitoring the boat traffic is essential. Automated moving object detection, tracking and object counting from a video surveillance are challenging tasks, but indispensable when quantitative models are needed for certain branches of management. Increasing technology in computer processors provides us with less computational time for large amount of data, which creates a great advantage to use computer-vision systems for object tracking and detection. Computer-vision system has wide application in the fields like traffic surveillance, security, criminology etc., all of which have stationary background. This paper introduces an algorithm to identify and count moving objects in a non-stationary background, namely boats navigating through Dalyan channel. The output accuracy of the algorithm is above 95% for daily boat count, which is quite high considering that the background is not stationary. The number of boats passing through Dalyan channel is investigated in daily and hourly basis for a period of 455 days. The daily time series, which has periodic character, clearly identifies the high tourist season and national holidays. Considering the seasonal trend of the boat traffic in Dalyan channel, seasonal management strategies can be developed for the sustainability of ecosystem in Dalyan.