Extracting roads from remote sensing images holds significant practical value across fields like urban planning, traffic management, and disaster monitoring. Current Convolutional Neural Network (CNN) methods, praised for their robust local feature learning enabled by inductive biases, deliver impressive results. However, they face challenges in capturing global context and accurately extracting the linear features of roads due to their localized receptive fields. To address these shortcomings of traditional methods, this paper proposes a novel parallel encoder architecture that integrates a CNN Encoder Module (CEM) with a Transformer Encoder Module (TEM). The integration combines the CEM’s strength in local feature extraction with the TEM’s ability to incorporate global context, achieving complementary advantages and overcoming limitations of both Transformers and CNNs. Furthermore, the architecture also includes a Linear Convolution Module (LCM), which uses linear convolutions tailored to the shape and distribution of roads. By capturing image features in four specific directions, the LCM significantly improves the model’s ability to detect and represent global and linear road features. Experimental results demonstrate that our proposed method achieves substantial improvements on the German-Street Dataset and the Massachusetts Roads Dataset, increasing the Intersection over Union (IoU) of road class by at least 3% and the overall F1 score by at least 2%.

In the field of agricultural robotics, robotic grippers play an indispensable role, directly influencing the rate of fruit damage and handling efficiency. Currently, traditional agricultural robotic grippers face challenges such as high damage rates and high requirements for position control. A robotic gripper for stable spherical fruit handling with high positional offset tolerance and a low fruit damage rate is proposed in this paper. It adopts a three-finger structure. A flexible active roller is configured at the end of each finger, allowing fruit translation with just a gentle touch. An integrated pressure sensor within the active roller further enhances the gripper’s compliance. To describe the effect of the gripper on the fruit, the interaction model was derived. Taking the tomato as a typical soft and fragile spherical fruit, three experiments were conducted to evaluate the performance of the proposed gripper. The experimental results demonstrated the handling capability of the gripper and the maximum graspable weight reached 2077 g. The average failure rate for the unilateral offset of 9 mm was only 1.33%, and for the bilateral offset of 6-6 mm was 4%, indicating the high positional offset tolerance performance and a low fruit damage rate of the gripper. The preliminary tomato-picking capability of the proposed gripper was also validated in a simplified laboratory scenario.

With the rapid development of artificial intelligence (AI), the Internet of Things (IoT), and robotics technology, the intelligent transformation of pig farms has become an inevitable trend in the livestock industry. In today’s large-scale pig farms, the traditional breeding methods are undergoing significant transformation due to the application of intelligent robotics technology. The robotic system is capable of performing autonomous inspection, precise feeding and environmental cleaning, which can effectively alleviate labor shortages on farms. It also shows great advantages in strengthening biosecurity, optimizing management processes and ensuring animal welfare. This paper systematically constructs the key technical framework of pig farm robots, including the basic support layer, the perception and execution layer, the intelligent processing layer, and the integrated application layer. On this basis, further analysis is conducted on the current application of robots in intelligent pig farms, covering the functional characteristics and technical implementations of inspection robots, cleaning robots, and feeding robots.

Traditional object detection methods for meat duck counting suffer from high manual costs, low image quality, and varying object sizes. To address these issues, this paper proposes FE-P Net, an image enhancement-based parallel density estimation network that integrates CNNs with Transformer models. FE-P Net employs a Laplacian pyramid to extract multi-scale features, effectively reducing the impact of low-resolution images on detection accuracy. Its parallel architecture combines convolutional operations with attention mechanisms, enabling the model to capture both global semantics and local details, thus enhancing its adaptability across diverse density scenarios. The Reconstructed Convolution Module is a crucial component that helps distinguish targets from backgrounds, significantly improving feature extraction accuracy. Validated on a meat duck counting dataset in breeding environments, FE-P Net achieved 96.46% accuracy in large-scale settings, demonstrating state-of-the-art performance. The model shows robustness across various densities, providing valuable insights for poultry counting methods in agricultural contexts.