DefinePK

Abstract

Cucumber is one of the major crops in Indian agrarian society, and it is affected by various diseases such as Downy Mildew. Monitoring the crop field's condition might help evade the disease, which is costly and time-consuming. Therefore, an economically intelligent farming system requires for disease monitoring. The grading of the disease can be recognized depending on the distribution of chlorophyll content in a leaf. However, previous grading techniques lead to an erroneous framework due to the inequitable statistics of real-time images' healthy and unhealthy pixel ordination. Hence, an optimized Deep Learning (DL) model is proposed according to the grading of the disease. The proposed DL model provides a training accuracy of 94.82% and a validation accuracy of 84.15%. The model also tested over 300 leaves with different grades of diseases captured randomly in an uncontrolled environment, and was found to be 90% accurate, compared to over 69% by visual identification of experts. A decision support system built on the proposed technology's instantaneous image capture and prediction capabilities is a huge help to farmers and agriculturists in understanding the state of the field and responding to such circumstances.