INTERNATIONAL RESEARCH JOURNAL OF ENGINEERING AND TECHNOLOGY (IRJET) E-ISSN: 2395-0056 VOLUME: 08 ISSUE: 01 | JAN 2021 WWW.IRJET.NET P-ISSN: 2395-0072 © 2021, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1080 A NOVEL APPROACH ON DISEASE AND SEVERITY DETECTION OF CROP AND PRDEICTION OF PESTICIDES USING MATLAB Gayatri Rallabandi 1 , Prof. Namrata Dewangan 2 , Prof. Dr. Pankaj Mishra 3 1 (M-Tech Scoholar), Department of Digital Electronics, Rungta college of Engineering and Technology, Kohka Road kurud Bhilai (C.G) 2 Assistant Professor, Department of Electronics and Telecommunication, Rungta college of Engineering and Technology, Kohka Road kurud Bhilai (C.G) 3 Professor, Department of Electronics and Telecommunication, Rungta college of Engineering and Technology, Kohka Road kurud Bhilai (C.G) -----------------------------------------------------------------------***----------------------------------------------------------------------- Abstract—Identification of plant diseases is important to avoid losses in yield and quantity of agricultural products. The study of plant diseases means the study of blind specimens found on the plant. Health monitoring and disease detection at plants is important for sustainable agriculture. Plant diseases are very difficult to monitor manually. It requires tremendous work, plant disease expertise and even high processing time. Therefore, image processing is used to detect plant diseases. Diagnosis includes stages such as image acquisition, image pre- processing, image segmentation, feature extraction, and classification. These methods are discussed in this paper detection of plant diseases using images of their leaves. Some segmentation and characteristic extraction algorithms used to detect plant disease are also discussed in this paper.. Key words- Deep learning, KNN, Plant diseases detection. I. INTRODUCTION The Indian economy depends on agricultural production. More than 70% of rural households are dependent on agriculture. Agriculture accounts for about 17% of total GDP [1] and provides employment to more than 60% of the population. Therefore the detection of plant diseases plays an important role in the agricultural field. Indian agriculture is made up of many crops such as rice, wheat. Indian farmers also grow sugarcane, seed oil, potatoes and non-food items such as coffee, tea, cotton, rubber. All these plants grow in strength of leaves and roots. There are factors that lead to various plant leaf diseases, which damage the plants and will eventually affect the world economy. This significant loss can be avoided by early detection of plant diseases. Accurate diagnosis of plant diseases is needed to strengthen the agricultural sector and our country's economy. Various diseases kill the leaves on the plant. Farmers find it very difficult to identify these diseases, which they are unable to detect in those crops due to lack of knowledge about those diseases. Biomedical is one of the fields for diagnosing plant diseases. Nowadays in the middle of this field, photo processing methods are suitable, efficient and reliable field for diagnosing diseases with the help of images of plant leaves. Farmers need quick and effective methods to diagnose all time-saving plant diseases. These programs can reduce efforts and the use of pesticides. In order to measure agricultural yields different ideas are suggested by scientists with the help of laboratory and plant diagnostic programs. The paper we have presented here researches different types of plant diseases and disease diagnostic techniques by different researchers. Objective The main objective of this project is to design a software tool to identify the crop disease by processing its leaf image, sending it to arboriculturist and receiving remedies. The underlying objectives are explained as follows: i. To apply image processing techniques to obtain affected portion of the crop and extraction of consequential feature values. ii. To perform comparison of extracted values with sample values to identify and classify the disease using various classifier algorithms. iii. To integrate and compare results of various classifier algorithms. iv. To predict the necessary control measures to cure the disease without any environmental and economic damage. II. LITERATURE SURVEY Yuanyuan Shao [16] discussed many features and genetic algorithm BP neural network. The Otsu method is used for partitioning and subtraction. According to real-time tobacco disease can be detected by the mobile customer and the server can make diagnoses of user-downloaded diseases. Here Otsu's method was used to rule out the local disease. The genetic algorithm can reduce training times and improve