The diverse visual features of tobacco leaves during various curing stages are influenced by multiple factors such as the origin of the tobacco and the environment of the curing room, making precise identification challenging with single features or models. To address this issue, this study proposes a Bayesian Mixture Expert Recognition Model for Tobacco Leaf Curing Stages based on feature fusion. First, deep learning models (ResNet34, MobileNetV2, EfficientNetb0) are utilized to extract deep features and traditional features positively correlated with curing stages from a constructed tobacco leaf image dataset. Various feature fusion methods (concatenate fusion, scaled fusion, adaptive gated fusion) are employed to construct multi-level feature representations. Next, different feature fusion methods of the same model are optimized to select the best-performing model as the foundational model for ensemble learning. Finally, Bayesian optimization is applied to integrate three optimized models, and comparisons are made with voting and weighted averaging methods. The proposed model achieves a recognition accuracy of 93.96% on the test set, with other performance metrics surpassing those of the base models. This research efficiently captures and robustly recognizes the complex dynamic visual features of the tobacco curing process through the integration of diverse features, adaptive adjustments, and expert collaboration mechanisms, thereby enhancing the system’s adaptability and interpretability in complex environments. This provides strong support for the intelligent upgrading of the tobacco industry.

Due to the constraints of the tobacco leaf curing environment and computational resources, current image classification models struggle to balance recognition accuracy and computational efficiency, making practical deployment challenging. To address this issue, this study proposes the development of a lightweight classification network model for recognizing tobacco leaf curing stages (TCSRNet). Firstly, the model utilizes an Inception structure with parallel convolutional branches to capture features at different receptive fields, thereby better adapting to the appearance variations of tobacco leaves at different curing stages. Secondly, the incorporation of Ghost modules significantly reduces the model’s computational complexity and parameter count through parameter sharing, enabling efficient recognition of tobacco leaf curing stages. Lastly, the design of the Multi-scale Adaptive Attention Module (MAAM) enhances the model’s perception of key visual information in images, emphasizing distinctive features such as leaf texture and color, which further improves the model’s accuracy and robustness. On the constructed tobacco leaf curing stage dataset (with color images sized 224×224 pixels), TCSRNet achieves a classification accuracy of 90.35% with 158.136 MFLOPs and 1.749M parameters. Compared to models such as ResNet34, GhostNet, ShuffleNetV2×1.5, EfficientNet-b0, MobileViT-xs, MobileNetV2, MobileNetV3-large, and MobileNetV3-small, TCSRNet demonstrates superior performance in terms of accuracy, FLOPs, and parameter count. Furthermore, when evaluated on the public V2 Plant Seedlings dataset, TCSRNet maintains an impressive accuracy of 97.15% compared to other advanced network models. This research advances the development of lightweight models for recognizing tobacco leaf curing stages, providing theoretical support for smart tobacco curing technologies and injecting new momentum into the digital transformation of the tobacco industry.

Vision is the most important way for an unmanned picking or plant protection robot to navigate an external environment. To achieve intelligent picking or plant protection, it is essential to obtain target location information. A new marker-controlled watershed (MCW-D) algorithm is proposed for object segmentation. By analyzing the shortcomings of the watershed algorithm and the characteristics of objects, the proposed MCW-D method mainly solves three problems. First, it reduces the influence of shadow and other factors on image color information. Based on histogram specification, secondary mapping is used to reduce the effects of lighting. Second, marker images are selected. All points with markers need to be located in the target object. The hue feature of long jujubes and trees is used as the marker image. Third, a mask image is acquired, which requires a clear boundary between the target and the background. An adaptive angle rotation based on an energy-driven approach is designed to find large differences between the target and the background. In a natural environment, the proposed MCW-D method respectively achieves segmentation accuracies of 94.7% and 93.2% on a jujube dataset and a tree dataset, which exceed the accuracies of widely used machine learning methods. These results promote the development of the forest and fruit economies.

It is important for the autonomous system to understand environmental information. For the autonomous system, it is desirable to have a strong generalization ability to deal with different complex environmental information, as well as have high accuracy and quick inference speed. Network ensemble architecture is a good choice to improve network performance. However, it is unsuitable for real-time applications on the autonomous system. To tackle this problem, a new neural network ensemble named partial-shared ensemble network (PSENet) is presented. PSENet changes network ensemble architecture from parallel architecture to scatter architecture and merges multiple component networks together to accelerate the inference speed. To make component networks independent of each other, a training method is designed to train the network ensemble architecture. Experiments on Camvid and CIFAR-10 reveal that PSENet achieves quick inference speed while maintaining the ability of ensemble learning. In the real world, PSENet is deployed on the unmanned system and deals with vision tasks such as semantic segmentation and environmental prediction in different fields.