Abstract
|
|
Webcams are available which tape the Panama Canal locks and waterways around the clock. The program developed seeks to automatically identify the presence of boats in a webcam image along the canal path. With this information, a person or organization could determine what kind of traffic the canal sees on a day-to-day basis.
|
Introduction
|
|
Hundreds of boats pass through the Panama Canal every day. Identification of these boats helps determine what kind of boats traverse the canal as well as how traffic flows through the canal on an hour-by-hour basis. A program that can automatically track boats can provide a wealth of information with minimal manpower.
|
Methods
|
|
This method used a moving average to compare an image to a number of past images, comparing the difference between the image and the average to the standard deviation. The program then searches for regions of pixels in the threshold image by counting adjacent pixels then highlighting those regions in the original image. The image below shows a threshold image for the green channel.
|
Results
|
|
Large boats are always found, but smaller boats are harder to find. Also, sudden changes in lighting conditions will cause mis-marked regions. Image 1 shows proper identification, while Image 2 shows mis-marked regions.
|
Privacy
|
|
People can track private vessels through the canals using these webcam images. Although records exist from payment for passage through the canal, publicly available images provide more details on time and speeds of boats.
|
References
|
|
Camera: https://www.pancanal.com/eng/photo/camera-java.html (choose the centennial bridge camera)
Project Paper (PDF)
|
