Enhancement of Crop Images Using Image Fusion Method in Transform Domain

Author Name(s): Dr. Vijaya Shandilya, Dr. Shyamkant S.Munje
Author Email: vkshandilya@rediffmail.com

Abstract

Computer vision and/or Image processing is the vast field and very much applicative in almost all sectors. We require reliable and accurate data from the images. The information which is extracted from the processed images is used in various domains. Here we chose agriculture domain where images are captured of different crops and then processed those images using image processing algorithm to improve the visual quality of the crop images. Once the clear image is available researcher from agriculture industry can identify problems associated with the crops. We are applying fusion technique to the set of crop images using complex wavelet transform (CWT) from the family of Discrete Wavelet Transform (DWT) implemented using MATLAB. This method enhanced the overall outcome of the fused image in terms of spectral and spatial characteristics of the image.

Keywords:  Image Fusion, CWT, DWT, MATLAB.

Introduction

The field of image processing has the wide range of applications where the outcome of processing in some desired form is given to the end user of the system. Image fusion is one of the method of image processing only the case is it involves multiple images whereas other processing mostly wok on single image. Fusion use images from single sensors to multiple sensors. It fuses the number of images using some fusion methodology to come up with accurate and reliable results [1][2]. Sometime information from single sensor is not clearly visible or absent, we can use images from different sensors to perceive this information. There are various application area were image fusion is prominently working such as [3], Military and law enforcement, Medical imaging, Robotic, Agriculture – crop analysis, crop disease identification, Remote sensing and so on. The proposed method work for most of the applications; where the images are multi-focus images, multi-view images and multi-modal images. Fusion is performed when the input images are not clear and user/ analyst are not able to provide information from it. It is very much helpful in the blurred images as we have number of images of same scene we can obtain its clear view by fusing partial blurred images or any combinations. Here we have chosen its most demanding application i.e. agriculture because it has direct benefit to the farmers and the society. Image Fusion is the process which provides clear and informative view of the field or crop under observation from number of images which are submitted as an input. Wide ranges of sensors are available nowadays which will provide crop images in different wavelengths. We can use those sensors to collect images and perform some processing and finally extract useful data from it. This data ultimately assist to agricultural scientist to identify certain problem or situation and finally serve the farmer and society. Manual analysis of these images is the cumbersome job. So we are applying fusion algorithm. This paper elaborates the method for fusion of number of images into one composite image which is accurate, reliable and full of information. The algorithm is based on complex wavelet transform (CWT) from the family of Discrete Wavelet Transform (DWT) implemented using MATLAB. This method enhanced the overall outcome of the fusion in terms of spectral and spatial characteristics of the image.

 

Conclusion

The proposed method can effectively use the complementary and redundant information of source images. The evaluation parameters demonstrate that the fused image preserve spectral as well as spatial characteristics of the source images. The method use strong edge features to enhance the final fused image. The proposed method also possesses property of perfect reconstruction which yield good fusion result compare to classical fusion method such as PCA [9]. Once the enhanced and clear image is available researcher from agriculture industry can perform various tasks on it. They can analyze crop health, crop density and coverage; stages of plant growth, various diseases associated with the crops and can provide solutions to them

932 total views, no views today

Download PDF File

About the author: admin