AI strategy for pest & disease detection on tomato plants

Abstract

India is an agricultural country and most of the people, wherein about 70% depends on agriculture. So, disease detection in plants is very important. Tomato is one of the strongly grown and widely used crops. There are many types of tomato diseases and pests, in which the pathology of which is complex. Crop diseases are a major threat to crop production, but their identification remains difficult in many parts of India due to the lack of the necessary infrastructure. It is very prone for attacks by aphids, whiteflies, Thrips. It is difficult and error-prone to simply rely on manual identification in a large open area. Recent advances in computer vision made possible by deep learning has made the way for automatic disease detection. To monitor the health of the tomato crops in acres of land where we cannot monitor the output of each sensor individually, AI is used to increase the yield and quality of crops using a Convolution Neural Networks (CNN), k-means clustering, and acoustic emission.

INTRODUCTION

India has vast area, but the current status of agriculture management is not sufficient to provide everything to the population, which can be problematic. The solution to this issue is the practice of monitoring and protecting the crops in the open land farming. Automation system is the technical approach in which the farmers in the rural areas are benefited by automatic monitoring and controlling of pests and protecting the crops. It replaces the direct supervision of the human. Here, AI is used to monitor the health of the tomato crops in large acres of land where we cannot monitor the output of each sensor individually with the help of Convolution Neural Networks and K means clustering and thus increasing the yield and quality of the crops.

Save your time!
We can take care of your essay

• Proper editing and formatting
• Free revision, title page, and bibliography
• Flexible prices and money-back guarantee

Place an order

The development and growth of crop depends on the temperature and humidity. The controlling and monitoring of open land parameters play vital role in overall development of plant. The objective of our project is to design a simple, efficient “Arduino‟ based system for automation of open land. The project features monitoring, recording and controlling the values of temperature, humidity and soil moisture inside the open land. The Arduino used is a highly compact, durable and easily available. The values of temperature and soil moisture are continuously communicated through various sensors to the Arduino. Also proper design, selection, construction and the management of the open land using sensors would augur well to the growth of a crop.

LITERATURE SURVEY

Susperrangi, Carlos Rubio and Libor Lenza presented the development and comparison of two different approaches for vision based automated pest detection and identification using learning strategies [1]. Santhosh Adhikari and Er.Saban Kumar presented the classification and detection of the plant’s diseases automatically especially for the tomato plants [2]. Christina Mueller Blenkle and Sascha Kirchner designed a system in which the position and sound of hidden insects are also detected, but the settlement sound of grain can be mistaken for insect sound [3]. K.Narsimha Reddy, B.Polaiah and N.Madhu presented the overview of different classification techniques [4]. Preetha Rajan, Radhakrishnan B presented the study of various image processing techniques and applications for pest identification and plant disease detection [5].

PROPOSED WORK

Owing to the inaccurate prediction results which is obtained from use of thermal and image sensor we propose to introduce acoustic sensors. This system is basically proposed only for large areas in acres to increase the crop yield without getting affected by pests. The system is trained to read the sensor output and the prediction is done through machine learning which improves the accuracy. The use of AI reduces the work of labors.

a) Arduino Uno

Arduino is a single-board microcontroller, intended to make the application of interactive items or environment further useful. It involves the whole lot to support the microcontroller; without problems connect it to a laptop with a USB cable or power it with an ac to dc adapter or battery to get began out. The Uno differs from all previous boards in that it does no longer use the FTDI USB to- serial using drive. It has 14 digital input/output pins (of which 6 can be used as PWM outputs), 6 analog inputs, a USB connection, a vigor jack, a reset button and more. It includes everything needed to aid the microcontroller; conveniently join it to a laptop with a USB cable or vigor it with a AC-to-DC adapter or battery to get started.

b) Acoustic Sensor

Literally acoustic or sound sensor is used to detect the sound. It is a small board that combines a microphone and some processing circuitry. The sound detector not only provides audio output, but also a binary indication of the presence of sound, and an analog representation of its amplitude. Early detection of pests in images is very crucial for effective management of pest control .but by using this acoustic sensor we can efficiently detect the pest at early stage.

Fig (c): Acoustic Sensor

c) Soil moisture sensor

Soil Moisture sensor is a sensor which detects the moisture substance of the soil. At the point when the soil is dry, the current won't pass through it thus it will go about as open circuit. Subsequently the yield is said to be most extreme. At the point when the soil is wet, the current will go from one terminal to the next and the circuit is said to be short and the yield will be zero. The sensor is metal covered to make the proficiency high. The scope of detecting is likewise high.

d) Temperature sensor

The LM35 is one kind of commonly used temperature sensor that can be used to measure temperature with an electrical o/p comparative to the temperature (in °C). It can measure temperature more correctly compared with a thermistor. This sensor generates a high output voltage than thermocouples and may not need that the output voltage is amplified. The LM35 has an output voltage that is proportional to the Celsius temperature. The scale factor is .01V/°C. The LM35 does not need any exterior calibration and maintains an exactness of +/-0.4°C at room temperature and +/-0.8°C over a range of 0°C to +100°C.One more significance of this sensor is that it draws just 60 micro amps from its supply and acquires a low self-heating capacity. The LM35 temperature sensor available in many different packages like T0-46 metal can transistor-like package, TO-92 plastic transistor-like package, 8-lead surface mount SO-8 small outline package.

CONCLUSION

This project is used to automate the open land with early detection of pest using acoustic sensor. This method yields more crops than the existing method with the improvement in their quality. The soil moisture, temperature, light intensity are measured and automatically controlled with IOT and AI technology. It has been interfaced with arduino, thus the open land has been automated.

REFERENCE

1. Learning strategies for pest detection and identification on tomato plants for autonomous scouting robots using internal databases. Altor Gutierrez, Ander Ansuategi, Loreto Susperrangi, Hindawi, Journal of Sensors, Volume 2019.
2. Tomato plant disease detection system using image processing. Santhosh Adhikari, Er.Saban Kumar, KEC Conference, Volume 1, September 27, 2018.
3. Plant leaf diseases detection using image processing technique. K.Narsimha Reddy, B.Polaiah, N.Madhu IOSR Journal of Electronics and Communication, Volume 12, Issue 3, Ver2, May-June 2017.
4. Plant leaf diseases detection using image processing technique. Preetha Rajan, Radhakrishnan B, International Journal of Computer Science and Network, Volume 5, Issue 1, Feb 2016.
5. A new approach to acoustic insect detection in grain storage. Christina Mueller Blenkle,Sascha Kirchner,Isabell Szallies,Cornel Adler. Germany 2018.
6. Improved efficiency of Insect pest control system by SSPA. Phanupong saeung Samran santalunai Thanaset 2018 IEEE, Thailand. Classification And Prediction Of Brinjal Leaf Diseases Through Image Segmentation. International journals of computer trends and technology, April 2017.
7. An Advanced Method for Chilli plant disease using image processing. Dipak,swapnil R. Kurkute,Pallavi S.Sonar, Antono, ICEST 2017.
8. Ecology and control of brinjal insect pest from Kolhapur region, India. Sathe,patil,devkar, Govail,Biolifejournal Vol 4,Issue 1,2016.
9. Chilli Leaf Curl Virus disease:a serious threat for chilli cultivation, Journal of Plant Diseases and Protection 20th January 2018.