In this paper, a novel Siamese network-based lightweight framework is proposed for automatic tomato leaf disease recognition. This framework achieves the highest accuracy of 96.97% on the tomato subset obtained from the PlantVillage dataset and 95.48% on the Taiwan tomato leaf disease dataset. Experimental results further confirm that the proposed framework is effective with imbalanced and small data. The backbone network integrated with this framework is lightweight with approximately 2.9629 million trainable parameters, which is second to SqueezeNet and significantly lower than other lightweight deep networks. Automatic tomato disease recognition from leaf images is vital to avoid crop losses by applying control measures on time. Even though recent deep learning-based tomato disease recognition methods with classical training procedures showed promising recognition results, they demand large labeled data and involve expensive training. The traditional deep learning models proposed for tomato disease recognition also consume high memory and storage because of a high number of parameters. While lightweight networks overcome some of these issues to a certain extent, they continue to show low performance and struggle to handle imbalanced data.

The detection of animals, in particular wildlife animals, is important to monitor their distribution for the conservation of animal life and to address a broad range of questions of animal researchers, such as in the study of ecosystem function and behavioural ecology, analyse the growth and development of animals, understand population dynamics, and discover the factors influencing animal movements. Researchers use camera traps that are activated when an animal enters their field, allowing them to collect millions of images of animals without disturbing them. Machine learning methods, particularly convolutional networks, have quickly risen to prominence as the preferred way for detecting and recognizing animals in camera trap images. This paper examines the major deep learning ideas relevant to the detection and recognition of wildlife animals, as well as the contributions to the field, the majority of which have been published recently. We survey the use of deep learning techniques for automated animal recognition, segmentation, and detection and provide a concise analysis and comparison of these approaches. The open challenges and prospective research directions are discussed.

Forest ecosystems are critical components of Earth’s biodiversity and play vital roles in climate regulation and carbon sequestration. They face increasing threats from deforestation, wildfires, and other anthropogenic activities. Timely detection and monitoring of changes in forest landscapes pose significant challenges for government agencies. To address these challenges, we propose a novel pipeline by refining the U-Net design, including employing two different schemata of early fusion networks and a Siam network architecture capable of processing RGB images specifically designed to identify high-risk areas in forest ecosystems through change detection across different time frames in the same location. It annotates ground truth change maps in such time frames using an encoder–decoder approach with the help of an enhanced feature learning and attention mechanism. Our proposed pipeline, integrated with ResNeSt blocks and SE attention techniques, achieved impressive results in our newly created forest cover change dataset. The evaluation metrics reveal a Dice score of 39.03%, a kappa score of 35.13%, an F1-score of 42.84%, and an overall accuracy of 94.37%. Notably, our approach significantly outperformed multitasking model approaches in the ONERA dataset, boasting a precision of 53.32%, a Dice score of 59.97%, and an overall accuracy of 97.82%. Furthermore, it surpassed multitasking models in the HRSCD dataset, even without utilizing land cover maps, achieving a Dice score of 44.62%, a kappa score of 11.97%, and an overall accuracy of 98.44%. Although the proposed model had a lower F1-score than other methods, other performance metrics highlight its effectiveness in timely detection and forest landscape monitoring, advancing deep learning techniques in this field.

Many real-world agricultural applications, such as automatic pest recognition, benefit from lightweight deep learning (DL) architectures due to their reduced computational complexity, enabling deployment on resource-constrained devices. However, this paradigm shift comes at the cost of model performance, significantly limiting its extensive use. Traditional data-centric approaches for improving model performance, such as using large training datasets, are often unsuitable for the agricultural domain due to limited labeled data and high data collection costs. On the other hand, architectural improvements, such as attention mechanisms, have demonstrated the potential to enhance the performance of lightweight DL architectures. However, improper integration can lead to increased complexity and diminished performance. To address this challenge, this study proposes a novel mechanism to systematically determine the optimal integration configuration of popular attention techniques with the MobileNet lightweight DL architecture. The proposed method is evaluated on four variants of two benchmark plant pest datasets (D1_5,869 and D1₅₀₀, D2₁₅₉₉, and D2₅₄₅) and the best integration configurations are reported along with their results. The Bottleneck Attention Module (BAM) attention mechanism, integrated into 12 different layers of MobileNetV2 (BAM12), demonstrated superior performance on D1₅₈₆₉ and D1₅₀₀, and D2₁₅₉₉ and D2₅₄₅, while integrating BAM into eight layers yielded higher accuracy on D2₁₅₉₉. As a result, a comparison with the MobileNet baseline demonstrates that the careful integration of attention mechanisms significantly improves performance.

In the process of quality analysis of grains, number counting is one of the key steps for agricultural production. Traditional manual grain counting methods are time-consuming and subject to human error, and automated methods have the potential to improve accuracy and save time. In this study, we aimed to develop an image analysis-based method to automatically quantify the number of grains in a quicker manner. The 576 grain images were collected manually, and labelImg tagging tool used to annotate to generate a text file with their respective positions by drawing bounding boxes manually. The datasets consist were separated into three groups: training, validation, and test. For object detection, the YOLOv5, YOLOv4, and YOLOv3 algorithms represent cutting-edge deep learning frameworks. They replace the tedious and error-prone manual counting process by precisely identifying and counting grains in images obtained from agricultural fields. This technique helps to increase grain counting's precision and effectiveness. We believe this method will be extremely beneficial in guiding the development of high throughput systems for counting the number of grains in other crops as it performs well with a wide range of backgrounds, picture sizes, grain sizes, as well as various quantities of grain crowding. When compared to the other two approaches, YOLOv4 performed well in terms of accuracy, speed, and robustness (97.65%), demonstrating that the suggested strategy is competitive with other cutting-edge deep networkst.

Following better agricultural practices is the key to catering for the ever-increasing food demand. While new technologies have been adapted over the years, there is still a need for effective plant disease recognition systems because of the existence of harmful plant diseases that can spread rapidly. Effective and early recognition of plant diseases is vital to minimize the damage to crops and hence can save the farmers from potential loss. It is also important for many countries to maintain economic stability, especially for the countries that completely rely on agriculture. In the past, many traditional and deep learning-based approaches have been proposed for plant disease recognition. While traditional approaches need insightful domain expertise, deep learning-based approaches require large sets of labeled data. Further, most of the existing methods fail to meet benchmark performances in terms of recognition accuracy. Therefore, in this study, a novel deep hybrid architecture is proposed to perform plant disease recognition from plant leave images. The Google Inception and ResNet architectures are utilized as the core networks to construct the proposed network. The proposed framework is evaluated on a newly constructed dataset with large sample size. The comparative analysis reveals that the proposed approach can outperform other state-of-the-art deep networks.