Cybernetics approaches in intelligent systems for crops disease detection with the aid of IoT

Detection of crop diseases is imperative for agriculture to be sustainable. Automated crop disease detection is a major issue in the current agricultural industry due to its cluttered background. Internet of Things (IoT) has gained immense interest in the past decade, as it accumulates a high level of contextual information to identify crop diseases. This study paper presents a novel method based on Taylor-Water Wave Optimization-based Generative Adversarial Network (Taylor-WWO-based GAN) to identify diseases in the agricultural industry. In this method, the IoT nodes sense the plant leaves, and the sensed data are transmitted to the Base Station (BS) using Fractional Gravitational Gray Wolf Optimization. This technique selects the optimal path for data transmission. After performing IoT routing, crop diseases are recognized at the BS. For detecting crop disease, the input image acquired from the IoT routing phase is then forwarded to the next step, that is, preprocessing, to improve the quality of the image for further processing. Then, Segmentation Network (SegNet) is adapted to segment the images, and extraction of significant features is performed using the acquired segments. The extracted features are adapted by the GAN, which is trained by Taylor-WWO. The proposed Taylor-WWO is newly devised by integrating the Taylor series and WWO algorithms. The proposed Taylor-WWO-based GAN showed improved performance with a maximum accuracy of 91.6%, maximum sensitivity of 89.3%, and maximum specificity of 92.3% in comparison with existing methods.

Automated summary

The study introduces a novel method based on Taylor-Water Wave Optimization-based Generative Adversarial Network (Taylor-WWO-based GAN) for identifying crop diseases in the agricultural industry. This method utilizes IoT nodes for sensing plant leaves, data transmission to a Base Station, image segmentation, feature extraction, and disease recognition. Experimental results demonstrate improved performance compared to existing methods.