International Journal of Pure and Applied Mathematics

Transcription

1 Volume 118 No , ISSN: (printed version); ISSN: (on-line version) url: ijpam.eu Weed Remover In Agricultural Field Through Image Processing Kaarthik K, Assistant Professor, Department of Electronics and Communication Engineering, M.Kumarasamy College of Engineering(Autonomous), Thalavapalayam, Karur, Tamilnadu, India. kaarthikk.ece@mkce.ac.in Vivek C, Associate Professor, Department of Electronics and Communication Engineering, M.Kumarasamy College of Engineering(Autonomous), Thalavapalayam, Karur, Tamilnadu, India. vivekc.ece@mkce.ac.in Abstract Agriculture is the back bone of the entire world. In olden days,more than 70% of the people are depend on Agriculture and related works. But now the theme changed the people are preferring the white collar jobs and profit motive made them to refuse the agricultural work. In other end the demand of food products are increased due to growing population. Inorder to fulfill the needs of the food and the other agriculture products the Automation is the right choice. Not only in the field of agriculture, the automation plays a vital role in each and every field to produce Quality output with reduced manpower. In the field of agriculture the man power is need much when compared to other fields, from ploughing to harvesting. In this process the removal of weeds is a tedious job and it requires more man power. Till date the weed removal can't be automated due to automation process will affect the main crops too. Later came the usage of deadly poisons called herbicides plays a drastic level in the removal of weeds and it also provides a good result at initial position but later the weeds dominate the field. Inorder to reduce the weeds the usage herbicide level increased day by day. Due to the usage of Herbicides causes damage to the land, the crops, Humans and to the Environment. So, in this project we have put into practice a some process to trim down the usage of the Herbicides by spraying herbicides only in the weeds. By this process the usage of herbicide level can be reduced. In this process the weed removal can done with the process of Implementation of Image processing using MAT Lab which detects the weed through the image which we took from the field. Keywords Image Processing; Plant Kingdom; Weed detection; Sprayer. I. INTRODUCTION In our country Agriculture is Traditional job. The each and every person of our country will came from Agricultural background. In those days all the members of the family depend only on Agriculture or Agriculture related works. They all will work in the field to get the yield. From the Germination till the Harvesting all work in the field and all the works were done manually. In those days the cultivation of goods will be much enough to fulfill the family needs and the Agriculture was done as a Service Motive one. The cultivating Land was also small and rest of land was used for Grassing Land for cattle's which was helpful to the Agriculture by Providing Manure and for Transportation of Goods. Later with Advancement in technology and Green Revolution induced the farmers to do the Agriculture in Profit motive results in increase of cultivating land. In other hand the younger generation prefer for other jobs. Then the usage of Natural Manure was replaced Artificial Fertilizers. Then they finally results in Lagging of Man power in the fields and increased cultivatable lands caused some Technology usage in Agriculture. But still now there were no technology is used to remove the weed. There were usage of some deadly poisons called Herbicides was introduced to remove the weeds [5]. The Herbicides were used by Spraying them in the Entire Field. They cause adverse impacts to the crops and to the Environment also, which results in some health problems to Humans who uses those Cultivatable Products. Inorder to reduce the usage of Herbicides the Image Processing were introduced. By the process of Image processing the Herbicides were been allowed to spray the chemicals only to the weeds. This process of spraying the Herbicides only to the weeds by analyzing the entire field with the help of Image processing. 393

2 II. IMPACT ANALYSIS OF WEED A. Weed Detection The weed can be a unwanted plant which is present in the field. They can cause some damage to the main crop and they can take the nutrition's from them. Inorder to get high yield the weed must been removed. They can be obtained all over the field and they are with different size and differ in Edge frequency, Boundary, Shape too[10]. The Weed can be controlled by removing them Manually or by Introducing some chemicals called Herbicides. If the Weed is not removed they spread seeds which acts as seed bank for many years. The weed can be classified according to the shape and edge frequencies as three types[4,7,9]. They are, 1) Weed with Narrow Leaves (Less Edge Frequency) Fig. 3. Weed with High Edge Frequency B. Methods of Weed Detection The weeds can be present in any areas of the field. But they are mostly present in between the crops and between the rows. In this paper the Weed between the Crops and Weed between the Rows are considered. 1) Inter Row Weed Detection In the process of the Inter Row Weed Detection the weeds present in between the rows of the crops are been considered and the image processing technique is introduced to spray herbicides upon them[2]. Fig. 1. Weed with Less Edge Frequency 2) Weed with Cluster Leaves (Medium Edge Frequency) Row 1 Weed Between Rows Row 2 Fig. 2. Weed with Medium Edge Frequency 3) Weed with Wide Leaves (High Edge Frequency) Fig. 4. Weed Detection Between Rows 2) Inter Crop Weed Detection In the Inter crop Weed detection the weeds present in between intervals between two crops. Those weeds are removed by sensing them through column weed detection process If the crops are planted uniformly throughout the field. But If the crops are planted randomly or spread irregularly in the field they are done by the process of image processing[8]. 394

3 START Crop 2 Weed Between the Crops Pick K Cluster centers Crop 1 Assign each pixel in the Image to the Cluster Fig. 5. Weed Detection Between Crops C. Image Segmentation An Image segmentation is the process of partition the image into a set of multiple segments as pixels. This segmentation process is to represent the image easier to analyze at a glance. The lines and curves which are typically called as Objects and boundaries respectively are represented through Image segmentation by assigning each and every pixel which share certain ocular Characteristics. The result obtained will be set of pixel segments which collectively represents the entire image.[1] The Image segmentation process will provide good quality output result. They undergone the Various steps which involves the Pre Processing step. The pre processing step undergone the process of De-Noising and Image Enhancement. The De-Noising process is provided with a non linear filter called Rank Filter. They can able to identify the weed by the data which we provided earlier as Shape, Edge, Boundary, Object etc.,. 1) Clustering The Clustering is a type of image segmentation process which can partition the into clusters. The clusters can selected on the basis of some conditions called manual selection or random selection which provides the absolute difference between the pixel and a cluster moments. The discrepancy provides the details of Pixel colour, Intensity, Texture and location factors[3, 10-12, 14]. The Certain segment is utilized to differentiate the colours as their group in the form of Pixels which results in the process named as Clustering. The colours obtained only in two as Black and white i.e. Greyscale to provide the projected the Output image. The colour green which represents the Crop and Weed once it undergone the Image Segmentation process will replicate as white and the rest of the parts will be replicated as Black part in the particular image. Re-Computed Cluster centre by Averaging all Pixel in the Cluster Whether there is convergence YES END NO Fig. 6. Flow Chart of Image Segmentation Process using Clustering The Process of technique named K-Means clustering has been used in the process of Texture Analysis. In the process of K-Means Clustering the algorithm is used to classify the pixels into number of classes which results in feature extraction which can be allowed to reduce the sum of squares that the distances has to obtained as the Objects and the respective cluster called as corresponding cluster. The corresponding cluster can also be called as the class centroid. The operation results in differentiate the pictures through the partitioning technique in which a single image can be converted into the clusters of number in four in which the affected part of leaf can be identified in ease. when the leaf shows the symptoms that is has been infected by more than one disease[3]. The process of clustering algorithm can be represented using Flow chart in Fig. 6. III. PROPOSED WORK In our Paper the weed can be detected by taking several images. The process will take 25 frames of pictures per second. The each image can be taken by 0.04 sec. and total of first eight images were taken in 0.3 sec. finally set of frames of images can be taken by 0.3 sec. and then they consider for set other frames and consider the next image. 395

4 During this process they will consider Edge detection and colour segmentation process. They will analyze each and every crop with their intensity colour, Intensity, edge, Size etc., are been obtained as output. The image can obtain the process of two operation called as the colour segmentation and the edge segmentation. They can be used to identify the exact identification of the edge and the veins of the particular crop which replicates in the form of white in colour where the rest of the parts are represented as Black in colour. After the process of Edge detection and Colour segmentation it undergone the process of Filtering. The Filtering can be a type of process in which each and every crop can be identified and the their gain value, trade-off's, edge, frequency of crop and weed can be identified and their threshold values can obtained[6,11]. Fig. 9. Output image after edge detection Fig. 7. Input Image taken for Weed Detection In the above mentioned crop the weed have high edge frequency on comparison with the crop. Then their edge frequency is high. In this paper the Corn crop is considered, which is a narrow leaf crop and their edge frequency is less when compared with weed. Inorder to Identify the edge frequency of the given image the for loop can be used to determine the clear cluster image of weed with number of blocks can be divided the number of blocks to determine the edges and the process results in identification of accurate weed by the process which is turned off after 900 different views. The same procedure is continued by considering the exact clear plant image to obtain the edge frequency mostly of 100. In our proposed work in order to identify the weeds we have considered the threshold value as 500. Being the threshold value is of 500 the image consideration will be of 400 * 500 pixels value as size. so the 100 blocks can be divided in the size of 40 * 50 pixels. In the proposed work we have considered two parameters or loop structures are used named as For loop and If loop. The For loop is used to count the number of pixels. The If loop is used to identify the Threshold values. The white Pixels are converted from all pixels of that particular block through the For loop process once if the Edge frequency value is higher when compared to the Threshold value which results in obtaining the accurate weed affected blocks and the rest of the blocks remains unchanged. The stimulation is carried out with the help of MAT Lab software and the obtained result is as follows: Fig. 8. Output Image after Colour Segmentation 396

5 V. CONCLUSION In this paper we have Proposed a method to identify the weeds through image processing. Then we will provide the information about the weed to the Automatic sprayer to spray Herbicides in that particular areas. So the usage of Herbicides can be reduced in big extent. But the only drawback of this process is that if there will be increase in weed type that is more than two weeds then the process will not be applicable for that and also if there will be any weed with less edge frequency cannot be Applicable. If there will be any field with crop and weed will has same edge frequency then the process will not be applicable for that and which might been identified. Fig. 10. Final output after filtering showing the weed affected areas as white blocks IV. AUTOMATIC SPRAYER The main objective of our paper is to reduce the usage of Herbicides and man power in the agriculture. In this paper the image from filtering can be considered by concentrating the weed block numbers. The sprayer we use may be a robotic hand or a set of sprinklers or motors etc.. The output of the image processing can be interfaced with the Robots or other external devices and they can be interfaced by the "Arduino uno" microcontroller. The Flowchart represents the Process undergone in the Automatic sprayer. Identification of Weed Affected Blocks Calculation of Delay for the Motor to reach those Blocks Giving Delay Input for all Motors Movement of Motor to that block and switch on Sprinkler to spray Herbicide Bring back the Motor to Initial Position Fig. 11. Flow Chart Operation of Automatic Sprayer REFERENCES [1] Murali Krishnan; Jabert.G., "Pest Control in Agricultural Plantations Using Image Processing", IOSR Journal of Electronics and Communication Engineering, Volume 6, Issue 4, May - Jun. 2013, pp [2] Mrs. Latha, A Poojith, B V Amarnath Reddy, G Vittal Kumar, "Image Processing in Agriculture", International Journal of Innovative Research In Electrical, Electronics, Instrumentation And Control Engineering, Vol. 2, Issue 6, June [3] Sabah Bashir, Navdeep Sharma : Remote Area Plant Disease Detection Using Image Processing, IOSR Journal of Electronics and Communication Engineering (IOSRJECE), ISSN : , Volume 2, Issue 6, Sep-Oct 2012, pp [4] Hossein Nejati, Zohreh Azimifar, Mohsen Zamani, "Using Fast Fourier Transform for weed detection in corn fields", IEEE International Conference on Systems, Man and Cybernetics, 2008 [5] Anup Vibhute, S K Bodhe,"Applications of Image Processing in Agriculture: A survey", International Journal of Computer Applications, Vol. 52, Issue 2, August [6] S. Thaiparnit and J. Jakkree Srinonchat,"Apply image processing technique to determine the correlation chlorophyll", 6th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology, Thailand, May 6 9, 2009, pp [7] J. I. Arribas, G. V. Sanchez-Ferrero, G. Ruiz-Ruiz, and J. Gomez-Gil, Leaf classification in sunflower crops by computer vision and neural networks, Computers and Electronics in Agriculture, Vol. 78, Issue 1, August 2011, pp [8] X. Pengyun and L. Jigang, "Computer assistance image processing spores counting", 2009 International Asia Conference on Informatics in Control, Automation and Robotics, February 1,2, 2009, pp [9] M. Baker, B. Carpenter, and A. Shafi, "Works in Progress: A pluggable architecture for high-performance Java messaging", IEEE Distributed Systems Online, Vol. 6, Issue 10, [10] M. Anitha, K. Kaarthik, "Analysis of nutrient requirement of crops using its leaf", Journal of Chemical and Pharmaceutical Sciences, ISSN No.: , 8, December 2016, pp [11] A. Meunkaewjinda, P. Kumsawat, K. Attakitmongcol, and A. Srikaew, "Grape leaf disease detection from color imagery using hybrid intelligent system", 2008 Proc. of ECTI-CON, pp [12] Yuvarani P 2012, "Image Denoising and Enhancement for Lung Cancer Detection Using Soft Computing Technique",IET Chennai 3rd International on Sustainable Energy and Intelligent Systems (SEISCON 2012) IEEE Digital Library, Online ISBN: DOI: /cp , Publisher.IET. [13] Yuvarani P and Maheswari S 2016, "Investigations of Various filters for lung Cancer CT Images", Journal of Chemical and Pharmaceutical Sciences, 8, Pages

6 [14] Y. Chen and X. Zhou, "Plant root image processing and analysis based on 2D scanner, in Proc. IEEE 5th International Conference on Bio- Inspired Computing: Theories and Application", Ghangsha, China, 2010, pp [15] Y. Al-Ohali, "Computer vision based date fruit grading system: design and implementation", Journal of King Saud University, Vol. 23, 2011, pp [16] C.Vivek, S.Autdithan, "Robust Analysis of the Rock Texture Image Based on the Boosting Classifier with Gabor Wavelet Features, Journal of Theoritical and Applied Information Technology", Vol. 69, Issue 3, 2014, pp [17] S.Palanivel Rajan, Diagnosis of Cardiovascular Diseases using Retinal Images through Vessel Segmentation Graph, Current Medical Imaging Reviews, Online ISSN: , ISSN: , Vol. : 13, DOI : / , [18] J.-X. Du, X.-F. Wang, and G.-J. Zhang, "Leaf shape based plant species recognition, Applied Mathematics and Computation", Vol. 185, Issue 2, February 2007, pp [19] P.R.Gomathi, V.Jamuna,"Application of Independent Component Analysis, International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, ISSN (Online): , Vol.3, special Issue 3, pp , [20] S.Palanivel Rajan, et.al., Visual and tag-based social image search based on hypergraph ranking method, IEEE Digital Library Xplore, ISBN : , INSPEC Accession Number : , DOI : /ICICES ,

7 399

8 400