Maydis leaf blight (MLB) of maize (Zea Mays L.), a serious fungal disease, is capable of causing up to 70% damage to the crop under severe conditions. Severity of diseases is considered as one of the important factors for proper crop management and overall crop yield. Therefore, it is quite essential to identify the disease at the earliest possible stage to overcome the yield loss. In this study, we created an image database of maize crop, MDSD (Maydis leaf blight Disease Severity Dataset), containing 1,760 digital images of MLB disease, collected from different agricultural fields and categorized into four groups viz. healthy, low, medium and high severity stages. Next, we proposed a lightweight convolutional neural network (CNN) to identify the severity stages of MLB disease. The proposed network is a simple CNN framework augmented with two modified Inception modules, making it a lightweight and efficient multi-scale feature extractor. The proposed network reported approx. 99.13% classification accuracy with the f1-score of 98.97% on the test images of MDSD. Furthermore, the class-wise accuracy levels were 100% for healthy samples, 98% for low severity samples and 99% for the medium and high severity samples. In addition to that, our network significantly outperforms the popular pretrained models, viz. VGG16, VGG19, InceptionV3, ResNet50, Xception, MobileNetV2, DenseNet121 and NASNetMobile for the MDSD image database. The experimental findings revealed that our proposed lightweight network is excellent in identifying the images of severity stages of MLB disease despite complicated background conditions.
Herbicides are the primary method for weed control in modern agriculture. Excessive herbicide usage harms both the soil and the environment. Large-scale weed identification by hand is impractical due to its expense and labour intensity, leading to suboptimal results. By combining selective and variable rate spraying, the smart herbicide sprayer effectively manages herbicide amounts. The accurate identification of weed type and density is crucial for its efficiency. In this study, we propose the ensemble of the lightweight deep neural networks, namely EfficientNetV2B0 and DenseNet121 for weed identification in the groundnut crop. These models are pretrained on ImageNet dataset. By using fuzzy rank-based fusion and three non-linear functions (tanH, exp and softplus), the proposed ensemble scheme combines the probabilistic scores of two base models. The final predictions on test samples are determined by the proposed ensemble technique, which differs from existing fusion schemes by incorporating the confidence in base model predictions. The proposed ensemble model has been trained and tested on the publicly available groundnut weed dataset (24,816 images of 16 weed classes), with tenfold cross validation technique for better performance results. Base models EfficientNetV2B0 and DenseNet121 have achieved the classification accuracy of 98.66 ± 0.39 and 98.92 ± 0.19 respectively. Furthermore, with the fuzzy rank-based fusion of these base models obtains an accuracy of 99.10 ± 0.18 with tenfold cross validation, thus increasing the overall performance and outperforms various state of the art models. The proposed fusion model of lightweight CNNs, can be embedded in IoT based agricultural system for weed identification in groundnut crop. This research contributes to sustainable agriculture through targeted herbicide application and reduced chemical overuse.