Agrobot- a Robot for Leaf Diseases Detection 1 2 Ruchita Bharat More , Prof

Research Article Volume 6 Issue No. 6

Agrobot- A Robot for Leaf Diseases Detection 1 2 Ruchita Bharat More , Prof. D. S. Bhosale ME Scholar1 Department of E&TC JSPM’s BSIOTR, Wagholi, Pune, India [email protected], [email protected]

Abstract: Diseases on plants cause significant damage and economic losses in crops. Subsequently, reduces the diseases on plant by early diagnosis results in substantial improvement in quality of the product. Incorrect diagnosis of disease and its severity leads to inappropriate use of pesticides. The goal of proposed system is to diagnose the disease by using image processing and artificial intelligence techniques on images of plant leaf. This system is divided into two phases, in first phase the plant is recognized on the basis of the features of leaf, it includes pre- processing of leaf images, and feature extraction followed by ANN based training and also classification for recognition of leaf. In second phase the classification of disease which is present in the leaf is done, this process includes K-Means based segmentation of defected area, feature removal of defected portion and the ANN based classification of disease. Then the disease grading is done on the basis of the amount of disease which is present in the leaf. Crops can be managed from early stage to harvest stage that is mature. It involves monitoring of plant diseases and also identification of disease, controlled irrigation, nutrition deficiency and controlled use of fertilizers and pesticides. Although the amount of remote sensing solutions is increasing, the availability and ground visibility during critical growth stages of crops continue to be major concerns. AGROBOT (a prototype) is a agricultural robot that can avoid challenges which is existing in big and complex satellite based solutions.

Keywords: Agriculture robot, Image Processing, Soil Moisture, Humidity Sensor, ANN.

I. INTRODUCTION give up the work of crop cultivation. There is therefore a need to identify these diseases at an early or superior stage and In Agriculture sector plants or crop cultivation have seen fast suggest solutions so that maximum harms can be avoided to development in both the quality and quantity of food increase crop yields. The applications like plant recognition, production, however, the presence of destructive insects and crop yield estimation, soil quality estimation etc. With the diseases on crops especially on leaves has hindered the quality existence of massive volume of plant species and their use in of agricultural goods. If the presence of pests on crops and various fields, the quality of agricultural products has become leaves is not checked properly and the timely solution is not a major issue in agriculture sector. Image processing provided then the quality and quantity of food production will technique such as machine vision system has been proven to be reduced, which results in upsurge in poverty, food be an effective automated technique. Image processing based insecurity and the mortality rate [6]. This severe effect can artificially intelligent computer vision techniques can reduce disturb any nation’s economy especially of those where 70% the computational time and as a result, the automated leaf of the inhabitants rely on the products from the agricultural disease detection can be made much faster. In future, the sector for their livelihood and endurance. One of the major farmer can obtain a consolidated view of the farm along with problems for agriculturists is to lessen or eradicate the growth decision support statistics for planning purposes. In the field of pests affecting crop yields. A pest is an organism that of agriculture digital image processing techniques have been spreads disease, causes damage or is a nuisance. The most established as an effective means for analyzing purposes in frequent pests that affect plants are aphids, fungus, gnats, various agricultural applications like plant recognition, crop flies, trips, slugs, snails, mites and Caterpillars. Pests lead to yield estimation, soil quality estimation etc. With the sporadic outbreaks of diseases, which lead to famine and food existence of massive volume of plant species and their use in shortage [3]. According to H. Al-Hiary et al [1] in most of the various fields, the quality of agricultural products has become countries farmers are used to detect pests manually through a major issue in agriculture sector. Image processing their observation of naked eyes, which requires continuous technique such as machine vision system has been proven to monitoring of the crop stem and leaves, which is a difficult be an effective automated technique. Image processing based labor intensive, inaccurate and expensive task for large farms. artificially intelligent computer vision techniques can reduce Further the early detection of diseases on plants is really the computational time and as a result, the automated leaf required as a very small number of diseased leaves can spread disease detection can be made much faster. the infection to the whole batch of fruits and vegetables and thus affects further storage and sales of agriculture products. This effect of plant diseases are very destructive as a lot of farmers were discouraged to the point where some decided to

International Journal of Engineering Science and Computing, June 2016 7352 http://ijesc.org/ A. Block Diagram

Fig.2. Defected Leaf Fig.1. Block Diagram II. METHODOLOGY B. Working Principle A. ARM9 A. Crop selection using Image processing The LPC2119/LPC2129 are based on a 16/32 bit Here we are connecting WEB CAM to PC which has the ARM7TDMI-S™ CPU with real-time emulation and MATLAB software. The MATLAB will take snapshot of the embedded trace support, together with 128/256 kilobytes (kB) of the SOIL samples. Here using color and texture of the soil of embedded high speed memory. With a wide range of we can predict the type of crop which is best suited for additional serial communications interfaces, they are also particular soil. suited for communication gateways and protocol converters as well as many other general-purpose applications. B. Leaf disease detection using Image processing Here we are connecting WEB CAM to PC which has the B. LCD MATLAB software. The MATLAB will take snapshot of the LCD is used in a project to visualize the output of the of the leaf samples. Here using color and texture of the soil application. We have used 16x2 LCD which indicates 16 we can track and detect the leaf disease to the farmer and the columns and 2 rows. So, we can write 16 characters in each type of pesticide to be used to the farmer. line. So, total 32 characters we can display on 16x2 LCD.LCD is used to check the output of different modules which is interfaced with the controller. Thus LCD plays a C. Robotic Buggy very important role in a project it shows the output and also to Here we are designing a robotic vehicle on which the webcam debug the system module wise in case of system failure which is mounted. We are going to program the buggy to stop in will rectify the problem. front of the plant. As soon as the buggy stops the MATLAB will take a snap of the plant using a WEBCAM. On C. Webcam MATLAB we are doing image processing and detect any A webcam is a video camera that feeds or streams its image in disease of the plant. If detected, then the buggy will start real time to or through a computer to computer network. spraying the selected pesticide on to the plant. When sent to a remote location, it will save the video stream or may be viewed. Unlike an IP camera (which connects D.MATLAB based GUI using Ethernet or Wi-Fi), Their most popular use is the We are designing an MATLAB based server which will establishment of video links, permitting computers to act receive the data from the µC. Depending on the reading the as videophones or videoconference stations.. Webcams are MATLAB will display these readings on GUI and also it will known for their low manufacturing cost and flexibility, send an SMS to the farmer informing him about the crop making them the lowest cost form of video telephony. selection and the status of plant growth.

International Journal of Engineering Science and Computing, June 2016 7353 http://ijesc.org/ D. Solenoid Valve the sensor like temperature sensor, humidity sensor and soil A solenoid valve is an electromechanically operated valve. If moisture sensor. Here water bottle used to supply the water to that force is sufficient to open and close the valve, then the plant in the soil after detecting the moisture. For that a direct acting solenoid valve is possible. An approximate purpose we are using the water motor and also two relay to relationship between the required solenoid force Fs, the fluid drive that motor. pressure P, and the orifice area A for a direct acting solenoid We are using Matlab for coding and to capture the image value is:Fs=PA=2πd^2/4 there is application on android phone which is called as IP WEBCAM. Database of healthy plant and diseased plant E. Temperature Sensor where already created in MATLAB. Communication of Temperature sensor is used to sense the temperature. We have Webcam to Matlab is done through wireless network like used a Temperature sensor called LM35. So, irrespective of hotspot. Then it will display the Temperature and moisture as the application to which it is used, it gives the reading of the shown is fig 4 and humidity s shown in fig 5 on LCD. After temperature. The LM35 is a temperature sensor, in which that robot will check the growth of the plant. According to when temperature increases the output voltage of LM 35 also that it will display the result like “Normal Growth” or increases. Temperature sensor is an analogue sensor and gives “Abnormal Growth” the output into form of analogue signal. This signal is feed to Then robot will capture the image of the plant by ADC which will convert it into digital form. using webcam and that image is sent to the Matlab for further operation. Matlab will compare the image with the database F. Humidity Sensor already saved in Matlab and if the plant have the disease then display the result like “Disease Detected” and if the plant does Humidity sensor is an analogue sensor and gives the output not have any type of disease then it will display the result like into form of analogue signal. This signal is feed to ADC “Healthy Plant”. Buggy will start and move forward to the which will convert it into digital form. When the signal is next plant for further operation. converted into analogue form, after that the microcontroller can process the digital humidity signal as per the application. This sensor gives the value of change in humidity in the atmosphere that is depends upon the application.

G. Soil Moisture Sensor Soil moisture sensors measure the water content in soil. The probe of soil moisture is made up of number of soil moisture sensors. The relation between the measured property and soil moisture must be calibrated and may vary depending on soil type.

H. RS 232 RS 232 is a serial communication cable used in the system. Here, the RS 232 is used in between microcontroller and the outside world for the serial communication. So it is a media used to communicate between microcontroller and the PC. In our project the RS232 serves the function to transfer the edited notice (or data) from PC (VB software) to the ARM9 controller, for the further operation of the system.

I. DC Motors Fig.3. Hardware Implementation DC motors are used to physically drive the application as per the requirement provided in software. The dc motor works on 12v.To drive a dc motor, we need a dc motor driver called L293D. This dc motor driver is capable of driving 2 dc motors at a time.

III. RESULTS The hardware of proposed system is in fig. 3, 4&5. Fig.3 shows the hardware implementation of project. It includes circuit in running mode. Fig 4 shows the readings of temperature and moisture by using temperature and soil moisture sensor. Fig 5 shows the reading of humidity by using humidity sensor. Figure6 Shows the full hardware implementation of the project. It is nothing but the Robotic buggy which contains circuit of ARM 9 controller and also all Fig.4. Reading of Temperature and Moisture

International Journal of Engineering Science and Computing, June 2016 7354 http://ijesc.org/ REFERENCES

[1] Aakanksha Rastogi, Ritika Arora, Shanu Sharma, “Leaf Disease Detection and Grading using Computer Vision Technology &Fuzzy Logic”

[2] Haiguang Wang, Guanlin Li, Zhanhong Ma, Xiaolong Li, “Image Recognition of Plant Diseases Based on Backpropagation Networks”

[3] Sai Kirthi Pilli1, Bharathiraja Nallathambi, Smith Jessy George, Vivek Diwanji , “eAGROBOT- A Robot for Early Crop Disease Detection using Image Processing”

[4] Fritz Brugger, “Mobile Applications in Agriculture”, Fig.5. Reading of Humidity Syngenta Foundation, Basel, Switzerland, 2011

[5] Pierre Sibiry Traoré, “The view from above” in ICT Update, a remote sensing scientist and GIS head at the (ICRISAT), 23 February 2010

[6] Lilienthal H, Ponomarev M, Schnug E 2004 Application of LASSIE to improve agricultural field experimentation. Landbauforsch Völkenrode 54(1):21-26 Online. Available: http://literatur.vti.bund.de/digbib_extern/bitv/zi032847.pdf

Fig.6 Full Hardware Implementation

IV. CONCLUSION

In the present scenario it is important to have an established approach for grading the defects on the plant leaves automatically. For this a system based on Machine Vision Technology and Artificial Neural Network (ANN) is of great use for automatically detecting the leaf plant as well as for leaf disease detection and grading. These systems are very helpful for agriculturist because it is efficient as compared to the manual method. The proposed system uses Euclidean distance technique and K means clustering technique for segmentation of image to segment the leaf area, disease area and background area of the input leaf image to calculate the percentage infection of the disease in the leaf and to grade them into various classes. These systems widely used to replace the old leaf diseases recognition technique and which is used by agricultural experts in identifying correct pesticide and its quantity to overcome the problem in an efficient and effective manner.

ACKNOWLEDGEMENT

I am thankful to Prof. D.S.Bhosale for his valuable support and guidance also for making this project successful.

International Journal of Engineering Science and Computing, June 2016 7355 http://ijesc.org/