11 II February 2023

https://doi.org/10.22214/ijraset.2023.48647
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 11 Issue II Feb 2023- Available at www.ijraset.com

Smart Agricultural Improvement using Digital

Image Processing
Saynee Paul1, Koushik Pal2, Chandrima Banerjee3, Suniket Pradhan4, Snehasish Bera5,
Aritra Das6, Sayani Saha7, Sneha Samanta8
1, 2, 3, 4, 5, 6, 7, 8
Department of Electronics & Communication Engineering, Guru Nanak Institute of Technology

Abstract: Image processing is an efficient tool that has been used for various field, agriculture is one of them. Sometimes expert
opinion or their ideas become more time consuming so by using this computer vision technology farmers can save times and
observe other major parameters like quality of crops, quantity of weeds, canopy etc. combination of image processing along with
availability of to communication network can be able to change the situation and save the time it’s also within affordable cost,
as image processing is a effective technique for analyzing parameter which we discuss in the above part. This paper is basically a
short survey on to discuss about some specific image processing techniques which can help the farmers to enhances the
agriculture practices, and also weed detection, gardening of fruits and also finding defects or brushes on fruits. By using all
these techniques the expertise as well as the farmer can say that it takes less time and the accuracy of the above mentioned
parameters is more productive than the traditional method. And application of image processing can also help to improve the
decision making for sorting of fruits, soaking, measurement of vegetation etc
Keywords: Digital image processing, Pesticides,Crops, Agriculture, Segmentation, Object detection, Image filtering, Background
extractions.

I. INTRODUCTION
Nowadays with the continuous increasing rate of population, reduction in cultivable land, weather change ,the demand for food is
also proportionally increasing, so to enhance the productivity and sustainability in the agriculture field , expertise invented many
new ways continuously. Their main intentions are to give minimal inputs for high production and also to save the time . For an
example Precision Agriculture is a method that is helping the farmers in meeting their needs.
Fertilizers, Pesticides, soaking and quality of crops are the major factors in agriculture. Maximum of the time expertise are required
to realize the problems which may be time taking and also there is a costlier issue in developing countries. So basically DIP is one
of the major techniques which can be used to measure all the parameters which is related to cultivation with certainty and economy.
This survey is mainly focus on many applications of DIP in various domains of agriculture field and also recent developments of
using computer-vision based applications. Basically image processing techniques entail five different basic processes such us
prepossessing, segmentation, object detection, image acquisition and classification.

Fig. 1 Flow diagram of the image processing method

A. Gray Scale Conversion
In 2014 Eerens et al, Jayas et al. in the year 2000 and Du and Sun in the year (2004) focused on gray scale conversion as an middle
step in food quality evaluation models. They are reported many applications calculating food items like fruits, meat, fishery e.t.c.
using of image processing techniques to various area of food quality assessment, the process is mainly pre-processing the images
that involves in gray scale conversion, after image acquisition.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 826
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 11 Issue II Feb 2023- Available at www.ijraset.com

Another work done by Wu et al. (2013),he was focused on the foreign fibers in such as cotton product images that increases
segmentation and background separation, use of gray level determination technique. In 2000 Jayas et al. basically show how to
classify the agriculture products by natural networks by using the image analysis techniques. using neural networks for classification
of the agricultural products. To categorizing of agricultural products the multi-layer neural network classifiers is the best way.

B. Image Object Analysis

Kelman and Linker proposed the shape and color analysis of the mature apples detection in tree images based on 3D convexity
analysis in 2014. It is the analysis of three-dimensional convex objects, the Golden Delicious apple variety orchard under natural
light conditions. The accuracy of the procedure correctness is about 94% in apples detection when the edges were identified using
Canny filter. The Segmentation of plant lets suffering with the problem of occlusion and testing with plants having 2, 3 and4 leaves
was Proposed by Pastrana and Rath in 2013. This method solved leaf complexities by ellipse approximation and found leaves
clusters using active shape models. Another study by Yang et al. shows integration of a system using digital camera and a personal
computer in précising decision-making in using herbicides in agricultural fields. This system processes the color images of the
agricultural fields the prediction in the reduction of herbicide use is around 15-64%.Another Study proposed by Chikushi et al in
1990 measures the cucumber root systems. This system obtained 98% accuracy in root length measurement with advantages that
there is no effect of randomness and direction of root images, and the system does not need any additional tool.

Fig. 2 pictures of image object analysis process

C. Feature Extraction
Weis and Gerhards predicted the attempts to detect weeds using image segmentation from images in 2009. They classified weeds
based on shape features and perfectly classified weeds into weeds classes. The results obtained were better than manual weed
sampling. The textural features like gray level co-occurrence matrix needed gray levels in an image; It was reported by Ehsanirad
and Kumar in 2010 who extracted these features for leaf recognition for plant classification. Other work by Jayas et al in 2000
studied morphological, Fourier descriptors, wavelet transforms, boundary chain codes, spatial moments, color and textural features
in the images and extraction techniques. In This study the further discussed features of classification techniques, including nearest
neighbor classifier, neural networks, and multi-layer neural network.

D. Image Filtering
In 2012 Sansao et al. proposed a unique approach by using Gabor filter. To get the assess in weed coverage percentage in images.
This image filtering method used for filtering crops regions to find the patches of weeds using excess green index images. Guijarro
et al. (2011) invented an automatic image segmentation algorithm , it is use to separate barley, cereal and corn crops including
weeds from the soil.

Fig. 3 Classification of different green plants and soil

In this above method the farmers can calculate the future improvements in the productiveness of this method by homomorphism.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 827
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 11 Issue II Feb 2023- Available at www.ijraset.com

E. Image Background Extraction

The main concept of Image background extraction is to minimal use of background and it’s prefer to extract it form the images. and
we get a result in non-uniform gray-level distribution which is between objects of interest and the image background, In 2004 Du
and Sun approaches various applications where while making the food products quality including pizza, cob and corn genetics, etc
background is not taken as into consideration. In the same way, Wu et al. (2013) removed background of the foreign fiber images
detected in cotton products. Sankaran et al. (2010) focused on finding the plant diseases and also future improvement by some
techniques like visible and infrared spectroscopy, hyper spectral imaging, fluorescence spectroscopy and imaging. By this above
study expertise could focus on the metabolic activity of the plants.

F. Binary Image Segmentation

In the year 2004 Du and Sun invented various food quality assessment techniques by separating defects and feebleness in the food
products using image segmentation. They described the productiveness of various segmentation algorithms applied for apple defects
detection, pizza sauce separation and detecting touching pistachio nuts. Another image processing technique invented by Puchalski
et al. (2008) which is find out the defects on the apple surfaces, by this method we can get the 96% perfectness in finding of bruises,
frost damages, and scabs from the combination from the images.

Fig. 4 Result of surface defect finding ,(a) Apples with one kind of surface defect.
(b) Apples with two kind of surface defects.

In 2008 by Tellaeche et al. using binary image segmentation to separate weeds from crops. And this method is mainly used in
pesticides management weed affected areas and this method is using by the farmer for decision making. Montalvo et al. in 2013
invented an automatic expert system by using binary image segmentation technique. It’s used for identify corps and weeds.
Precision agriculture is a well known technique in agriculture field. This process mainly helps farmers to select the proper seeds,
monitoring the disease of the fruits, fertilizer and pesticides usage. And more over many parts of the world specifically in rural areas
this kind of information are not easily asses sable and also it’s enough costly for the farmer’s. The process is mainly depends on
location specific data and also it includes of multiple image databases . This image processing technique is used to help in terms of
decision making by providing high resolution pictures.

II. CONCLUSIONS
Image processing is a well known technique which has proved as a efficient computer-vision system in agriculture field as well.
Some of the different range like Infrared, X-ray were very useful in terms of identify irrigated land mapping, measurement of
canopy etc. There are so many methods and also some new methods that are evaluating to assist and facilities in the farming
practices. it’s noticeable that all the approaches which were contribute the bigger goal of improve the global food production. , we
can conclude that image processing is the most conventional and effective tool that can be used for agriculture field with a huge
perfectness and provincial parameters.

REFERENCES
[1] Image analysis for apple defect detection was invented by Puchalski, J. Gorzelany, G. Zagula and G. Brusewitz. In the year 2008. detection, which is published
at Biosystems and Agricultura Engineering, 8, pp. 197–205, 2008.
[2] C. C. Yang, S. O. Prasher, J. A. Landry, J. Perret and H. S. Ramaswamy Identification of weeds with the help of image processing and they use fuzzy logic for
precision farming in the year 2000. At Canadian Agriculture Engineering.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 828
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 11 Issue II Feb 2023- Available at www.ijraset.com

[3] Methodology for Potatoes Defects Detection with Computer Vision was invented by J. Jin, J. Li, G. Liao, X. Yu and L. C. C. Viray in the year 2009 which
was published in the International Symposium on Information Processing (ISIP’09)
[4] In 2012 J. P. H. Sansao, M. C. Silva, L. A. Mozelli, F. A. C. Pinto and D. M. Queiroz using digital images, for Weed Mapping and it was published at
International Conference of Agricultural Engineering which was held in July 2012.
[5] A Review -Applications of Thermal Imaging in Agriculture written by Annamalai Manickavasagan, Digvir S. Jayas, Noel D.G. White, Jitendra Paliwal, at
the Canadian society for engineering in agriculture ,food and biological system.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 829