Intelligence has been considered as the major challenge in promoting economic potential and production efficiency of precision agriculture. In order to apply advanced deep-learning technology to complete various agricultural tasks in online and offline ways, a large number of crop vision datasets with domain-specific annotation are urgently needed. To encourage further progress in challenging realistic agricultural conditions, we present the CropDeep species classification and detection dataset, consisting of 31,147 images with over 49,000 annotated instances from 31 different classes. In contrast to existing vision datasets, images were collected with different cameras and equipment in greenhouses, captured in a wide variety of situations. It features visually similar species and periodic changes with more representative annotations, which have supported a stronger benchmark for deep-learning-based classification and detection. To further verify the application prospect, we provide extensive baseline experiments using state-of-the-art deep-learning classification and detection models. Results show that current deep-learning-based methods achieve well performance in classification accuracy over 99%. While current deep-learning methods achieve only 92% detection accuracy, illustrating the difficulty of the dataset and improvement room of state-of-the-art deep-learning models when applied to crops production and management. Specifically, we suggest that the YOLOv3 network has good potential application in agricultural detection tasks.

Food quality and safety issues occurred frequently in recent years, which have attracted more and more attention of social and international organizations. Considering the increased quality risk in the food supply chain, many researchers have applied various information technologies to develop real‐time risk identification and traceability systems (RITSs) for preferable food safety guarantee. This paper presents an innovative approach by utilizing the deep‐stacking network method for hazardous risk identification, which relies on massive multisource data monitored by the Internet of Things timely in the whole food supply chain. The aim of the proposed method is to help managers and operators in food enterprises to find accurate risk levels of food security in advance and to provide regulatory authorities and consumers with potential rules for better decision‐making, thereby maintaining the safety and sustainability of food product supply. The verification experiments show that the proposed method has the best performance in terms of prediction accuracy up to 97.62%, meanwhile achieves the appropriate model parameters only up to 211.26 megabytes. Moreover, the case analysis is implemented to illustrate the outperforming performance of the proposed method in risk level identification. It can effectively enhance the RITS ability for assuring food supply chain security and attaining multiple cooperation between regulators, enterprises, and consumers.

In emerging sensor network applications, localization requirement of applications and availability needs feasible algorithm with lower cost and higher accuracy. To localize the wireless sensor networks nodes accurately, we propose error-compensation based node distance estimation, which benefits from joint exploitation, at no cost, of the information provided by the adjacent areas and nodes. Under the given limitation, the localization problem is transformed into total least squares with inequality constraints (ICTLS) model with three estimated parameters, and the problem is solved by iteration of inequality constraints errors in variables. Simulation results show that the performance of our proposed algorithm is superior to DV-Hop algorithm and other improved algorithms in all considered scenarios.

In this paper, a novel semi-supervised segmentation framework based on a spot-divergence supervoxelization of multi-sensor fusion data is proposed for autonomous forest machine (AFMs) applications in complex environments. Given the multi-sensor measuring system, our framework addresses three successive steps: firstly, the relationship of multi-sensor coordinates is jointly calibrated to form higher-dimensional fusion data. Then, spot-divergence supervoxels representing the size-change property are given to produce feature vectors covering comprehensive information of multi-sensors at a time. Finally, the Gaussian density peak clustering is proposed to segment supervoxels into sematic objects in the semi-supervised way, which non-requires parameters preset in manual. It is demonstrated that the proposed framework achieves a balancing act both for supervoxel generation and sematic segmentation. Comparative experiments show that the well performance of segmenting various objects in terms of segmentation accuracy (F-score up to 95.6%) and operation time, which would improve intelligent capability of AFMs.

Wireless sensor networks (WSN) can be used in many fields. In wireless sensor networks, sensor nodes transmit data in multi hop mode. The large number of hops required by data transmission will lead to unbalanced energy consumption and large data transmission delay of the whole network, which greatly affects the invulnerability of the network. Therefore, an optimal deployment of heterogeneous nodes (ODHN) algorithm is proposed to enhance the invulnerability of the wireless sensor networks. The algorithm combines the advantages of DEEC (design of distributed energy efficient clustering) clustering algorithm and BAS (beetle antenna search) optimization algorithm to find the globally optimal deployment locations of heterogeneous nodes. Then, establish a shortcut to communicate with sink nodes through heterogeneous nodes. Besides, considering the practical deployment operation, we set the threshold of the mobile location of heterogeneous nodes, which greatly simplifies the deployment difficulty. Simulation results show that compared with traditional routing protocols, the proposed algorithm can make the network load more evenly, and effectively improve energy-utilization and the fault tolerance of the whole network, which can greatly improve the invulnerability of the wireless sensor networks.