|
|
| |
| Smart Security System Using Camera Trap Based on Background Subtraction |
Sudjadi, Maman Somantri, M Arfan, Imam Santoso, Sukiswo5, Agung Nugroho
|
Abstract:
In recent years efforts to increase fishery products as one of the export commodities implemented by various methods, either through intensification of cultivation or extensive by opening potential ponds. One source of such failure is the low supervision of the cultivation environment so that theft and sabotage cases increase each year. The case of theft and sabotage in aquaculture is very detrimental to farmers and increased in the period before the harvest. The crime is detrimental to the pond management. Control of ponds by placing the guards ineffective due to the location of ponds far and extreme temperatures. On the other hand, information technology is growing especially in the field of pattern recognition that can be used to detect crime in the pond. By utilizing the development of computer vision and human recognition technology by detecting motion patterns, intelligent surveillance detection systems can be implemented. In this research, the main issues inherent to camera trapping images automatic human detection on coastal aquaculture. Through several experiment the capacity of convolutional neural networks to automatize human behavior activity in camera-trap images. By using the internet of things architecture that supported the camera-trap device in the pond is expected to facilitate the manager in monitoring the condition of the ponds remotely so that the pond becomes safe. The method of recognizing the crime behavior with the background and Activity Pattern will make the introduction more precise and as expected. With the support of the internet of things architecture will help in real-time monitoring and have connectivity to several places at once.
|
|
|
| | |