This paper presents a security system for a remote farm. This research is ideal for locations where real time monitoring and notifications are paramount to your security wherein the image(s) detected through a surveillance system is processed by the newly advanced technology such as Open-source Computer Vision programming, Matlab programming and Artificial Intelligence Network.
The research is all about detection process through Open-Source Computer Vision programming wherein multiple object detection can be applied. Future recommendations for image processing and recognition can be used through Matlab programming technology and Artifical Intelligence Network Technology.
REVIEW OF RELATED LITERATURE AND STUDIES
Ambient Intelligence in Home Environments. The project is all about Computer Vision Tracking System. The message that comes from the warning was sent thru email or via alarm. For the part of human tracking, he uses optical flow method and differential motion analysis.
For the fire detection, he uses flame recognition method through video and implementation of Open Source Computer Vision (OpenCV) Library. He uses servo motors controlled via ActiveX coming from the serial port. The software consist of Graphical User Interface (GUI). Detection can be actuated in real time .
A. Computer Unit:
Any personal computer can be used in the system like desktop computer or notebook. The computer has at least 1.5 GHz processor, 300 MB RAM, 20 GB hard disk.
In this paper, it is suggested to use the thermal image camera. A thermal image camera allows to photographs when it is dark out. It is also an infrared camera that performs by exposing the charge coupled device (CCD) to the infrared light of the spectrum which cannot be seen to the naked eye.
C. Detection Part- Open Source Computer Vision (Open-CV) Programming Using Background Subtraction:
Open source computer vision library was based in C/C++ programming language. It was created and maintained by Intel. It has a feature of cross-platform, portable, and good for real-time applications. It is a library for real time image processing specifically for computer vision or machine vision. OpenCV can be used for Human-Computer Interaction (HCI), Object Identification, Segmentation, and Recognition Face Recognition; Gesture Recognition; Motion Tracking, Ego Motion, and Motion Understanding; Structure From Motion (SFM); and Mobile Robotics.
DATA AND RESULTS
The code was compiled and run using OpenCV and Visual Studio. Median blurr threshold adjustments were added to the code in order to improve the detection by ignoring background noises from sudden change in lighting.
ANALYSIS OF RESULTS
One known weakness of background subtraction comes from the fact that pixels are independently processed or computed. Even minor changes in the pixels within the frames are being considered as motion. Lighting condition greatly affect the efficiency of this algorithm. But “false positive” pixels can be cleaned up using image processing operations embedded in OpenCV such as cvErode(), cvDilate(), cvFloodFill(). These operations eliminate stray patches or pixels .
CONCLUSIONS AND RECOMMENDATIONS
A neural based face recognition system is found to be invariant to changes in illumination for background and illumination conditions. One way to overcome this is through neural network training. The feedforward back propagation do not have feedback connections, but errors are back propagated during training. During training, the net output is compared with the target value and the appropriate error is calculated.
The face recognition system whether human or animal creature is implemented using a Matlab software package. In this method we use feedforward back propagation neural network. It is an information processing system and has been developed as a generalization of the mathematical model of human recognition. The function of a neural network is to produce an output pattern for a given input patterns when presented.
Source: De La Salle University
Authors: Alexander C. Abad | Jose B. Lazaro Jr. | Elmer P. Dadios