An unique model for weed and paddy detection using regional convolutional neural networks

Abstract

Aim: In the agricultural field, weeds are grown irrespective of the required species, which spoils the growth of paddy plants. The presence of weeds is to be detected and should be classified in the earlier stage to improve the growth of species. This research work considers paddy cultivation and detection of weeds in the paddy field. Methods: The modelling of the automatic weed predictor model aids farmers in handling the weed coverage and scattering of weed in the agricultural field. Real-time data is collected from the agricultural region, and the images are provided as the input for the predictor model. Regional Convolutional Neural Networks (R-CNN) is proposed to segment the weed from the input images. Results: The model is proposed to address the segmentation problem by concurrent simulation of the task for object prediction. Simulation is carried out in a MATLAB environment. The performance of R-CNN is compared and evaluated with existing approaches like the conventional CNN model and other segmentation approaches. Conclusion: The proposed model gives better results when compared to other approaches.