Search

: In order to monitor illegal land use in real time, video surveillance technology was used to monitor the vulnerable areas of illegal land use. Excavator was one of the most important construction machinery in the engineering construction, an automatic real-time detection of excavator could provide important information for non-contact field monitoring of illegal land. But it was difficult to accurately detect the excavator due to the complex background, uneven illumination and partial occlusion in natural scene, This paper proposed a real-time excavator detection algorithm in natural scene based on the SSD-like (Single Shot Detector). Specifically, the lightweight network DDB (Depthwise Dense Block) was used as the basic network to extract shallow feature and fuse with high-level features in the excavator objection model to enhance the feature representation capability. Meanwhile, the BDM (Bottleneck Down-sampling Module) which was designed based on the lightweight network MobileNetV2 was used as the multi-scale feature extraction network to reduce the parameter quantity and computation. The data sets included 18 537 images of excavators with different shooting angles and natural scenes, 15 009 images were used as training set and 3 528 images were chosen as test set randomly. To enhance the diversity of training data, data set expansion methods such as rotation and image were adopted. Based on the Caffe deep learning framework, the proposed model in this paper was trained with end-to-end approximate joint methods and the model weight was fine-tuned by using SGD (Stochastic Gradient Descent) algorithm. The DDB module of the network was initialized with the weights pre-trained on the PASCAL VOC dataset, and the training time and resources were further reduced by transferring learning. Then the model pre-trained on the large data sets was transplanted to excavator object detection by transfer learning. The proposed method was transplanted and performed on the mainstream Jetson TX1 embedded hardware platform, and experiments on the actual data set of detecting excavator object at different angles of view and natural scenes. Experiment results showed that the parameter quantity and computational complexity of proposed model with BDM was reduced by 68.4% compared to SSD, the mAP (Mean Average Precision) of proposed method reached 90.6% on the testing set, which was 0.4% and 0.4% higher than that of SSD based on VGG16 basic net and MobileNetV2SSD based on MobileNetV2 basic net, respectively. The model size of propose method was 4.2 MB, which was about 1/25 and 1/4 of SSD and mobilenetv2ssd, respectively, and the time-consuming of each frame was 145.2 ms, which was 122.7% and 28.2% faster than SSD and MobileNetV2SSD, respectively. The proposed method not only had better generalization and robustness, but also can be better deployed on the embedded hardware platform which demonstrated the feasibility of real-time monitoring of the excavator at site of illegal land use. [ABSTRACT FROM AUTHOR]

From the problem of mad cow disease to date, the traceability system as effective means to ensure food safety has been introduced for nearly 30 years. Now, reducing breakage degree for traceability chain, enhancing traceability credibility, and improving early warning capacity for agri-food quality and safety, these has increasingly become hot topics of traceability system research, which is also urgent problems in traceability system application. Focusing on these issues, the agricultural products smart traceability system affected by new generation information technologies based on the relevant literature was summarized and analyzed. Firstly, the development stages of traceability system from 1.0 to 3.0 were presented. The features of the three stages were information recording for 1.0, data integration for 2.0, and intelligent decision-making for 3.0, respectively. Secondly, the relationship of new generation information technology including internet of things (IoT), big data, cloud computing, blockchain, artificial intelligence (AI) was described. The new generation of information technology is not only the vertical promotion of single technology, but also the horizontal penetration of technology integration. The development trend about big data, artificial intelligence and block chain were summarized. Finally, the relevant research on the issues of reducing the breakage degree for traceability with AI, improving early warning capability for agri-food quality and safety with big data, and enhancing the traceability credibility of the whole supply chain with blockchain were summarized. The orientation of in-depth study was put forward in the light of the technology development trend. For combing AI, traceability granularity should be improved from within and among supply chains through establishment some flexible and intelligent traceability models. For coming big data, prediction quality and optimization processing model should be enhanced through data mining and analysis from different levels. For coming blockchain, traceability block structure optimization, privacy protection, block chain consensus algorithm should be performed to meet the requirement agri-food. In fact, AI, big data, block chain and other technologies have different characteristics. In order to improve the intelligent level for traceability system, the different technologies should be integrated deeply. This paper provides a useful reference for grasping the development trend of traceability system, for knowing the research hot spot and for understanding the application bottleneck. [ABSTRACT FROM AUTHOR]

In weed control, using unmanned aerial vehicle (UAV) to obtain images, spraying specific pesticides according to different weed communities is an effective means of prevention and control. Sum-product networks is suitable for small embedded devices such as UAV. But it has many parameters, long training time, and more redundant nodes and subtrees in the image classification task, so that the recognition accuracy is not high. In response to these problems, this paper improved the learning process of traditional sum-product networks and used a mini-batch learning method to construct a network model through one pass of data. Its lightweight structure required less hardware resources and was more suitable for small embedded devices such as drones. It had reference significance for the subsequent spraying of pesticides by drones. For the input image, the light sum-product networks weed recognition model first used K-means clustering as the low-level feature extractor to obtain the feature dictionary, then downsampled the extracted features, and took the sampling features into mini-batches of data as input to train the light sum-product networks. Each category corresponds to an independent network structure, and the high-level features were extracted by internal nodes in the network structure. The probability values of the corresponding categories were output by the root nodes to identify weeds. The network structure was updated by comparing the correlation coefficients between variables. Bayesian moment matching was used to update the network parameters. To simplify the structure, when a product node had only one child, it was removed from the network, and its child nodes were connected to its parent node. Similarly, if a sum node was the last node of another sum node, then the child node was deleted and all its child nodes were promoted one layer up. This effectively reduced redundant edge branches and made the model structure lighter. Using this method, the average classification accuracy of soybean seedlings, grass weeds, broadleaf weeds and soils in UAV images was 99.5%, and the average sensitivity was 99.6%. And the model parameter quantity was only 33% of the traditional sum-product networks. The parameter quantity would increase with the input of the data flow. The amount of parameters was still much smaller than traditional convolutional neural networks AlexNet when using the larger data sets to construct the light sum-product networks. It showed that the model was suitable for larger data sets. The memory usage was reduced by 549 M compared to the traditional sum-product networks and was reduced by 1 072 M compared to the convolutional neural networks. The maximum average training time was reduced by 688.79 s compared to the traditional sum-product networks, which was much less than the convolutional neural networks. The experimental results showed that using the light sum-product network as the weed recognition model, the model parameters were less, the memory requirements were lower, and the training time was shorter without loss of precision. The shortcoming was that the data acquired by the UAV image in the previous stage needed to be processed in multiple steps. The data set itself relied on manual classification. Some images had different categories of overlap, and the misclassification of such images would increase when the features were extracted. However, by adjusting the classification threshold, the overall classification can achieve the desired results. The research can provide a reference for the use of light sum-product networks in weed recognition of UAV spraying pesticides. [ABSTRACT FROM AUTHOR]

The livability evaluation of living space was an important topic, which had important guiding significance for improving the livability of living space.This study aimed to explore the theoretical logical relationship between the multi-function of land use and livability of living space, and make use of the research about multi-function of land use for the optimization practice of livability of living space. So, this paper discussed the theoretical logical relationships between multiple functions of land use and livability of living space from the perspective of human needed products and services supplied by land use, and the evaluation index system was built based on the relationship and previous research. And then, the scheme of index spatialization was designed based on the technology of geographical information system(GIS) and remote sensing(RS). At the same time, multi-source data like remote sensing, survey, statistical and other spatialdataset were integrated by analytic hierarchy process (AHP) to quantify the spatial distribution of living space in ZhangbeiCounty, Hebei Province. The results showed that the livability of living space in ZhangbeiCounty presented the trend that the high value center was concentrated in the ZhangbeiTown and the subcenterwas distributed in Ertai-ShagouTown. The living space area above relatively strong levelwas only 2.42% of all county area.The spatial distribution characteristics of 3 types of land use were significantly difference;livingbearing function presented a trend like network distribution takingmain traffic trunk line as high value axis and ZhangbeiCounty as high value center. Living support function presented 2 groupsof high value centerscontaining the ZhangbeiTownand the north of YoulougouTown, and Ertai-shagouTown. The high value center of mental purification function was distributed in the ZhangbeiTown and the subcenterwas distributed in the area along the southwest of ZhangbeiTown to DaheTown and Ertai-shagouTown. There was a large difference in the distribution of livability between towns and villages. Most of the areas of livability space in town land were above the moderate level, especially, the area the livability level of which belonged to the higher and highest accounted for 78.37% of the total area of the town land, and meanwhile, accounted for 74.94% of the total area with higher and highest level of living space. This area was mainly distributed in the town of Zhangbei, Ertaiand Shagou, and YoulougouVillage. Most of the livability areas in rural land were under the moderate level, and the area oflower and lowest level accounted for 78.39% of the total area of the rural land, whatwas more, accounted for 95.32% of the total area with lower and lowest level of living space. This area was mainly concentered in the town ofHailiutu, Erquanjing, Dahe, Danjinghe, and Liangmianjing. The spatial difference of the functional elements of land use was the main reason to the difference in livability of living space. So, according to the above results, some conclusions can be drawn in this research. The villages, which containlarge lower and lowest livability space area, are the key areas of primary concern inthe implementation process of the“Revitalizing the countryside” strategy; what’s more, the theory and method of land use multi-function can be applied to the study of livability evaluation of living space. [ABSTRACT FROM AUTHOR]

Agricultural water saving is a main concern in long-term development in China. The combination of Internet of Things (IOT) technology and its applications of agriculture is an important driving force to promote the upgrading and transformation of traditional agriculture to modern agriculture. Focusing on the water need of open field plant (mainly refers to food crop), this paper summarized the development status and limitations when IOT technology was used in the key link of water saving applications. Three layers of IOT technology, including perception, transportation and application were discussed. Crop water need information perception layer was with a great importance in view of its fundamental role. This layer had 3 sections, small-scale soil moisture and crop water stress monitoring technology, medium-scale crop evapotranspiration quantification and estimation, and large-scale remote sensing including estimation of soil moisture, inversion and prediction for evapotranspiration were analyzed in detail. For the small-scale, different soil moisture detection methods were compared, including oven-dried method, neutron method, near-infrared reflectance method, tensiometer method and dielectric method. Typical equipments and sensors produced for the methods by different companies were also listed. In terms of crop water stress monitoring technology, 2 categories method such as tissue water status method and stomatal conductance method were listed and compared. The change from direct soil moisture monitoring to indirect crop water stress monitoring was presented as an new trend. For the medium-scale, 9 specific methods belonged to the 2 classification: measurement method and estimation method. In the aspect of large-scale, remote sensing was considered to be effective and economically feasible. Through spectrum detection method, a boarder range of crop evapotranspiration could be estimated and predicted, and it promoted the development of agriculture water saving from micro to macro. In the information transportation layer, environmental adaptability, communication reliability, cost and power consumption of the terminal equipment which was used in open filed scene were discussed. Meanwhile, the common wireless communication technologies were compared in terms of frequency, communication distance, data rate, security, etc. In the application layer, the advanced soil moisture monitoring and irrigation decision support system application status in China and abroad were analyzed, and the importance of public information service was pointed out. Based on the different characteristics of the 3 layer structure and Chinese agricultural develop strategy, the developing trend of water saving techniques, software service, association relationship, industrialization development mode for agriculture, and application pattern were also discussed under the changed background of fast development of new technologies of Big Data, Cloud Computing and Internet Plus technologies, etc. The roles and status of hardware used in application would change to software and its supporting service. It could help users to cut the cost investment. The revolution from causality to association promoted by the combination of Big Data and connection would be in favor of reasonable planning and effective utilization of water resource. Land reform policy in China would be an important driver for the changof traditional agriculture, and would make good use of agronomic technology, machinery technology, information technology and automation technology. Besides, suggestion was also made to technology supply based on actual demand, and the harmonious development of public service and customized service. In conclusion, 3 new strategies were proposed, including change from technology to service, from determination to prediction and from micro to macro. [ABSTRACT FROM AUTHOR]