Plant disease is the major threat to the productivity of the plants. Identification of the plant diseases is the key to prevent the losses in the productivity and quality of the yield. It is a very challenging task to identify diseases detection on the plant for sustainable agriculture, where it requires a tremendous amount of work, expertise in the plant disease, and also requires excessive processing time. Hence, image processing is used here for detection of diseases in multi-horticulture plants such as alternaria alternata, anthracnose, bacterial blight, and cercospora leaf spot and also addition with the healthy leaves. In the first stage, the leaf is classified as healthy or unhealthy using the KNN approach. In the second stage, they classify the unhealthy leaf using PNN, SVM, and the KNN approach. The features are like GLCM, Gabor, and color are used for classification purposes. Experimentation is conducted on the authors own dataset of 820 healthy and unhealthy leaves. The experimentation reveals that the fusion approach with PNN and SVM classifier outperforms KNN methods.
TopIntroduction
Indian economy is majorly depends upon the Agriculture with around 70% of the population earning their livelihood directly or indirectly from this sector as per the 2011 census (Government of India, 2017). Presently, India ranks second worldwide in horticulture farm produce. The economic contribution of agriculture to India's GDP (Gross Domestic Product) is steadily declining with the country's broad-based economic growth. Still, agriculture is demographically the broadest economic sector and plays a significant role in the overall socio-economic fabric of India (Government of India, 2017). Agriculture is a critical sector as it plays major part in the development of the Indian economy and food security for the vast population of the country. As India is a rain fed area, agriculture majorly depends on the climatic conditions, and due to the adverse effect of nature like excessive or deficit rainfall, extremely hot or dry weather which affects the growth stages of a plant, which will impact on the crop productivity. Current agriculture modernization practice has opened a new avenue due to the globalization and liberalization policies of this country (Pradnya A Joshi, 2015). Even though many farmers try to follow the modern agricultural practice they fail to achieve higher productivity due to various reasons. A special attention for development planning in the field of agriculture is needed which will eventually help the sector to sustain, hence a sustainable agriculture would be the choice.
Horticulture crops grown mainly for the purpose of food, materials, comfort and beauty i.e., flowers, fruits, vegetables and plants for ornament and fancy. These crops are continuously under the intervention of plants parasites and pests due to various pathogens attack. During the growth period of this plant it is attacked by a number of diseases from bacteria, fungal and virus. Bacteria grow by decaying the organic matter in soil and multiply themselves in the tissues of the plant and Fungi looks like thread vegetative growth, in the presence of moisture they germinate spores and produce infection as both of these are microorganisms. Virus diseases cause a streaking of the stem, which will kill the growing conditions that affect the growth of the plants (Mahlein, 2016). Some of the major diseases affecting the growth of the Alternaria Alternata, Anthracnose, Bacterial Blight and Cercospora Leaf Spot and also addition with the healthy leaves. From all these diseases lead to a reduction in productivity and downward in the profit to the farmer (Joshi, 2015)
Typically, detection of diseases in the plant leaves has been carried with the experience by farmer’s i.e. visual inspection of the cultivators and they apply some fertilizers or pesticides to overcome from that disease. The potential of detecting different diseases in the plant is desirable, where in the growth stages of plants may be affected simultaneously by many pathogens, such as bacteria, fungal and viruses (Mahlein, 2016). Nowadays, automatic detection of diseases attracts researchers from the different domains due to its benefit in monitoring large field and the consistency which it can provide (Gavhale and Gawande, 2016). A modern approach for detecting and analysis of leaf diseases are lacking to reach the expectation level. Even though using various detection, extraction and classification techniques in image processing are available but they have not been applied and proven on some of these vegetable plants under various conditions. Further farmer needs a better detection model to identify the different diseases affecting from the pathogens and measures to cure that disease so the application of pesticides, etc can be in precise quantity.