The plant disease recognition model based on deep learning has shown good performance potential. However, high complexity and nonlinearity lead to the low transparency and poor interpretability of such models. These limitations greatly limit the deployment and application of such models in field scenarios. To solve the above problems, we propose a dense caption generative model, Veg DenseCap. This model takes vegetable leaf images as input and uses object detection technology to locate abnormal parts of the leaf and identify the disease results. More importantly, it can describe the disease features it sees in natural language, and users can judge whether the relevant features are semantically consistent with human cognition based on these descriptions. First of all, a dataset containing Chinese feature description statements for images of 10 leaf diseases involving two vegetables (cucumber and tomato) was established. Secondly, Faster R-CNN was used as a disease detector to extract visual features of diseases, and LSTM was used as a language generator to generate description statements for disease features. Finally, the Convolutional Block Attention Module (CBAM) and the Focal Loss function were employed to overcome the imbalance between positive and negative samples and the weak performance of Faster R-CNN in obtaining key features. According to the test results, the Intersection-over-Union (IoU) and Meteor joint evaluation index of Veg-DenseCap achieved a mean Average Precision (mAP) of 88.0% on the dense captioning dataset of vegetable leaf disease images, which is 9.1% higher than that of the classical FCLN model. The automatically generated description statements are characterized by advantages of accurate feature description, correct grammar, and high diversity.

To enhance the transplantation effectiveness of vegetables and promptly formulate subsequent work strategies, it is imperative to study the recognition approach for transplanted seedlings. In the natural and complex environment, factors like background and sunlight often hinder accurate target recognition. To overcome these challenges, this study explores a lightweight yet efficient algorithm for recognizing cabbage transplantation states in natural settings. Initially, FasterNet is integrated as the backbone network in the YOLOv5s model, aiming to expedite convergence speed and bolster feature extraction capabilities. Secondly, the introduction of the GAM attention mechanism enhances the algorithm’s focus on cabbage seedlings. EIoU loss is incorporated to improve both network convergence speed and localization precision. Lastly, the model incorporates deformable convolution DCNV3, which further optimizes model parameters and attains a superior balance between accuracy and speed. Upon testing the refined YOLOv5s target detection algorithm, improvements were evident. When compared to the original model, the mean average precision (mAP) rose by 3.5 percentage points, recall increased by 1.7 percentage points, and detection speed witnessed an impressive boost of 52 FPS. This enhanced algorithm not only reduces model complexity but also elevates network performance. The method is expected to streamline transplantation quality measurements, minimize time and labor inputs, and elevate field transplantation quality surveys’ automation levels.

The kale crop is an important bulk vegetable, and automatic segmentation to recognize kale is fundamental for effective field management. However, complex backgrounds and texture-rich edge details make fine segmentation of kale difficult. To this end, we constructed a kale dataset in a real field scenario and proposed an UperNet semantic segmentation model with a Swin transformer as the backbone network and improved the model according to the growth characteristics of kale. Firstly, a channel attention module (CAM) is introduced into the Swin transformer module to improve the representation ability of the network and enhance the extraction of kale outer leaf and leaf bulb information; secondly, the extraction accuracy of kale target edges is improved in the decoding part by designing an attention refinement module (ARM); lastly, the uneven distribution of classes is solved by modifying the optimizer and loss function to solve the class distribution problem. The experimental results show that the improved model in this paper has excellent performance in feature extraction, and the average intersection and merger ratio (mIOU) of the improved kale segmentation can be up to 91.2%, and the average pixel accuracy (mPA) can be up to 95.2%, which is 2.1 percentage points and 4.7 percentage points higher than the original UperNet model, respectively, and it effectively improves the segmentation recognition of kale.

In recent years, remote sensing image processing technology has developed rapidly, and the variety of remote sensing images has increased. Solving a geographic problem often requires multiple remote sensing images to be used together. For an image processing analyst, it is difficult to become proficient in the image processing of multiple types of remote sensing images. Therefore, it is necessary to have multiple image processing analysts collaborate to solve geographic problems. However, as a result of the naturally large volumes of data and the computer resources they consume for analysis, remote sensing images present a barrier in the collaboration of multidisciplinary remote sensing undertakings and analysts. As a result, during the development of the collaborative analysis process, it is necessary to achieve the online processing and analysis of remote sensing images, as well as to standardize the online remote sensing image collaborative analysis process. To address the above issues, a hierarchical collaborative online processing and analysis framework was developed in this paper. This framework defined a clear collaborative analysis structure, and it identifies what kinds of online image processing and analysis activities participants can engage in to successfully conduct collaborative processes. In addition, a collaborative process construction model and an online remote sensing image processing analysis model were developed to assist participants in creating a standard collaborative online image processing and analysis process. In order to demonstrate the feasibility and effectiveness of the framework and model, this paper developed a collaborative online post-disaster assessment process that utilizes radar images and optical remote sensing images for a real forest fire event. This process was based on the BPMN2.0 and OGC dual standards. Based on the results, the proposed framework provides a hierarchical collaborative remote sensing image processing and analysis process with well-defined stages and activities to guide the participants’ mutual collaboration. Additionally, the proposed model can help participants to develop a standardized collaborative online image processing process in terms of process structure and information interactions.

Aiming at the problems of high cost of orchard agro-technique information service, and service channel is not smooth, a "IOT and cloud services" mode platform architecture of orchard information service was established, which is including networking aware transport layer, storage layer, service layer and application layer. Several platform construction key technologies are discussed. The platform has four sub platform, they are orchard data gathering, orchard production management, farm management and agricultural products circulation management providing information service for agro-techniques from production to circulation. Through the application analysis, the platform can not only gather resources provide a knowledge service for the promotion of agricultural technology, and provide technical support for the scientific management of orchard, also can effectively reduce the base information financial and human input.

Scenario aware computing is an important technology to realize precise information service in agricultural technology extension. This paper constructs a scene data fusion user perception of precision agricultural extension service. This paper established user classification model and multi situational awareness recommendation model based on multiple-factor, and realized the intelligent push services in personalized information. In comparison to common retrieval system which did not have the situational awareness and the intelligent push functions, the personalized information services based on situational awareness calculation was able to identify and predict the preferences of user interest, provided the psychological content for users. The correlation of personalized information services and resource push based on situational awareness calculation increased on average by 2.22 times, the accuracy rate increased by 1.34 times, meeting the user's personalized needs.