Most of the water quality indicators that affect results river assessment are gray and localized, thus correlation between can be calculated using analysis (GRA).However, GRA takes equal weighting for does not take into account indicators. Therefore, this paper proposes a method based on improved grey (ACGRA) andparticle swarm optimization multi-classification support vector machine (PSO-MSVM) assessing environment quality. Firstly, combination weights were Analytic Hierarchy Process...

Abstract Aiming at the problem that linear regression of traditional water quality prediction model is not robust enough and accuracy cannot be guaranteed in presence interference, this paper proposes an agricultural based on Momentum algorithm to optimize logistic (LRM algorithm). The uses quickly adjust misclassified samples. When object encounters a local optimum process falling, introduction momentum makes it easy for next update jump out with help last large gradient. In paper,...

Blueberry fruit phenotypes are crucial agronomic trait indicators in blueberry breeding, and the number of fruits within cluster, maturity, compactness important for evaluating harvesting methods yield. However, existing instance segmentation model cannot extract all these features. And due to complex field environment aggregated growth fruits, is difficult meet demand accurate automatic phenotype extraction environment. To solve above problems, a high-precision based on hybrid task cascade...

In order to realize the automatic fruit picking by robot in orchard environment, cherry target detection is carried out studying algorithm based on deep learning, so as identify and carry operation. However, most of existing methods are aimed at relatively sparse fruits cannot solve problem small dense fruits. this paper, a YOLOv5s-cherry model YOLOv5s was proposed for detection. By adjusting CSP structure BackBone Neck, deepening depth width network, improving ability network feature...

With the serious deterioration of water environment, accurate prediction quality changes has become a topic increasing concern. To further improve accuracy and stability generalization ability model, we propose new spatiotemporal forecast model to predict future quality. capture characteristics pollution data, three sites (station S1, station S2, S4) with highest temperature time series concentration correlation at experimental were first extracted 17,380 records collected each monitoring...

Due to the high cost of annotating dense fruit images, annotated target images are limited in some ripeness detection applications, which significantly restricts generalization ability small object networks complex environments. To address this issue, study proposes a self-supervised cherry algorithm based on multi-feature contrastive learning, consisting module and an module. The enhances features unlabeled through random augmentation, reducing interference from backgrounds. establishes...

HIGHLIGHTS Proposing a drinking-water data cleaning model with the combination of nonlinear partial differential equations and CART decision tree. Preprocessing by removing outlier elements not following normal distribution. missing modified classifier AdaBoost. Implementing proposed Big Data technology based on Hadoop architecture. Overall performance method competes other similar cutting-edge methods.

Accurate prediction of river water quality is great help for intelligent management and control the environment. Water data have complex correlations time series characteristics, but existing method in introduction LSTM model prediction, only forward processing, did not consider reverse processing data, ability to focus on local feature extraction different moment features enough. Therefore, this paper proposes a based AM-CNN-BiLSTM. First, CNN layer used extract dimensionality reduction...

Abstract Blockchain is an emerging distributed database technology. It has the characteristics of decentralization, non-tampering, traceability and final consistency. can solve data management problems in untrusted environments. Based on research analysis blockchain system, this article expounds application system environment. First, introduces concept classification from multiple aspects. Then it storage technology encryption adopted by detail. Finally, prospects today’s society are introduced.

With the rapid development of information technology, people's communication methods are also constantly improving. As most popular method, 5G has received widespread attention from all walks life. This article introduces history and some key technologies used in 5G, such as: Orthogonal frequency division multiplexing (OFDM) technology flexible frame design. It introduced important applications Internet Things, looked forward to prospects 6G.

Non relational database is a very popular and technology. It has the characteristics of large amount data, high concurrency, strong reading writing ability, flexible use, simple data structure good scalability. can solve problem cache reduce connection tables. Based on analysis performance non database, this paper compares several databases. Firstly, introduces concept classification database. Then read write commonly used non-relational databases MongDB Redis are compared in detail....

Intelligent agriculture has become the development trend of in future, and it a wide range research application scenarios. Using machine learning to complete basic tasks for people reality, this ability is also used vision. In order save time fruit picking process reduce cost labor, robot achieve automatic orchard environment. Cherry target detection algorithms based on deep are proposed identify pick cherries. However, most existing methods aimed at relatively sparse fruits cannot solve...

Aiming at the problem of low efficiency and large memory consumption in mobile robot path planning, this paper proposes an improved version RRT*Fixed Nodes( RRT*FN )algorithm based on adaptive threshold, named A-RRT*FN. The algorithm constructs a sampling tree two directions starting point, target point adaptively sets threshold according to environment both sides improve environmental adaptability search speed algorithm. Secondly, node elimination strategy is proposed quickly locate...

Combine the visual module with Delta parallel robot to set up sorting experiment platform based on technology. Matching results are sensitive environmental interference. In order eliminate influence such as illumination and noise matching process, traditional grayscale features replaced by edge features. The advantages of using that contains complete information about image there is less data be deal with. input images analyzed under multi-scale filtering in Canny algorithm frame extract...

The health degree of large distributed information system reflected from the business perspective is core index to measure stability system. It has significant meaning for fault diagnosis any To solve above problems, a knowledge representation method based on weighted proposed in this paper. Based method, an tree proposed. Then, model established. Finally, medical industry used verify that and are efficient practical, it can provide quantitative means In conclusion, paper find abnormal...

In order to study the Kinect and Unity3D combined with deeper application, based on use of identify body made a new attempt for binding Unity3D. Using get dynamic behavior in scene human body, achieve control Unity object scene. Build virtual platform. The test shows that recognition fast response, high accuracy. Otherwise, as platform is convenient, combination very suitable target system various industries.

A novel algorithm based on NURBS curve is addressed for acceleration/deceleration of CNC machine tools by combining quartic polynomial velocity planning and filtering, aiming at vibration caused sharp acceleration jerk. Its validity has been proved theoretically. By connecting a filter after the routine results, continuous jerk are further smoothed. Experiment shows that this simple effective to achieve smoothing jerk, thereby reducing achieving high speed, efficiency precision machining.

With the development of network society, individuals and enterprises generate massive data every day. However, traditional relational MySQL database has poor performance in storing processing data. In case huge amount data, may even crash, which obviously cannot meet needs storage. Obviously, with more emphasis on user-friendliness, storage application determine future enterprises. This paper discusses engine, high concurrency, architecture distributed cluster HBase, then compares...

Water quality prediction is a research hotspot in the field of ecological environment, which great significance to prevention water pollution and construction automatic monitoring network. The accuracy model results will affect scientificity correctness applied engineering projects, as well control measures. Firstly, background development trends models at home abroad are systematically introduced. Then, method based on machine learning mainly introduced, focusing time series method,...