Detection of vegetation in environmental repeat photography: A new algorithmic approach in data science

Environment change being one of the major issues in today’s world needs special attention of the researchers. With the advancement in computer vision researchers are equipped enough to come up with algorithms accomplishing automated system for environment monitoring. This paper proposes an algorithm which can be used to observe the change in vegetation utilizing the images of a particular site. This would help the environment experts to put on their efforts in a right direction and right place to improve the environment situation. The proposed algorithm registers the image so that comparison can be carried out in an accurate manner using single framework for all the images. Registration algorithm aligns the new images with the existing images available in the record of the same particular site by performing transformation. Registration process is followed by segmentation process which segments out the vegetation region from the image. A novel approach towards segmentation is proposed which works on the machine learning based algorithm. The algorithm performs classification between vegetation patches and non-vegetation patches which equips us to perform segmentation. The proposed algorithm showed promising results with F-measure of 85.36%. The segmentation result leads us to easy going calculation of vegetation index. Which can be used to make a vegetation record regarding particular site.