Abstract—The intelligent moving object detection has become one of the key research areas in the computer vision. Although, there are a lot of researches and methods have been proposed related to the intelligent moving object detection, and visual surveillance and intelligent recognition system. However, still it’s a great challenge of intelligent identification of moving object detection in the natural environment, due to the natural factors such as wind, sunlight, lighting and sudden illumination change which has been affecting the accuracy of moving object detection and intelligent recognition. For example, wind makes swaying trees and water rippling; sunlight makes shadows; lighting causes sudden change of light. To eliminate these problems, we have proposed a hybrid novel method based on Gaussian mixture model (GMM); Background subtraction; HSV color model; Feature extraction; and Neural networks. First, background is modeled with Gaussian Mixture Model (GMM), to eliminate the effect caused by the natural environment. Second, foreground image is extracted with background subtraction method. Third, the shadows of moving objects are detected and removed in HSV color model and morphological operation is done to get the clean foreground. That means detection is completed. Then, it is updated the background to adapt the dynamic background. After object detection, it is extracted shape features by using the Hu’s seven moment invariants of the training samples of the image data, which is used to train the back propagation neural network (BPNN) as input. Finally, we have done the intelligent identification process on the trained BPNN to recognize and distinguish the detected object whether it is human or pets. The algorithm can not only eliminate the effect of natural conditions, like wind, sunlight and lightning, but also automatically update the background when the illumination changes suddenly, or moving objects stop to move, or the background objects turn to move. The advantages of the proposed algorithm are accurately moving object detection, and the detection result is not affected by the body pose. The experimental results have shown that the proposed algorithm has good robustness and real-time performance in natural environment.
Index Terms—Feature extraction, HSV color model, intelligent identification, neural networks.
The authors are with mechanical engineering department in Tongji University, Shanghai, China (e-mail: 1793129@tongji.edu.cn).
Cite: Bhupendra Kumar Yadav and Jian Xiaogang, "An Algorithm for Intelligent Identification of Moving Objects in Natural Environment," International Journal of Machine Learning and Computing vol. 10, no. 5, pp. 637-647, 2020.
Copyright © 2020 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).