Search

In this study, a cotton disease diagnosis method that uses a combined algorithm of case-based reasoning (CBR) and fuzzy logic was designed and implemented. It focuses on the prevention, diagnosis and control of diseases affecting cotton production in China. Conventional methods of disease diagnosis are primarily based on CBR with reference to user-provided symptoms; however, in most cases, user-provided symptoms do not fully meet the requirements of CBR. To address this problem, fuzzy logic is incorporated into CBR to allow for more flexible and accurate models. With the help of CBR and fuzzy reasoning, three diagnostic results can be obtained by the cotton disease diagnosis system (CDDS) constructed in this study: success, success but not exact and failure. To verify the reliability of the CDDS and its ability to diagnose cotton diseases, its diagnostic accuracy and stability were analyzed and compared with the results obtained by the traditional expert scoring method. The analysis results reveal that the CDDS can achieve a high diagnostic success rate (above 90%) and better diagnostic stability than the traditional expert scoring method when at least four disease symptoms are input. The CDDS provides an independent and objective source of information to assist farmers in the diagnosis and prevention of cotton diseases.

Automatic visual inspection (AVI) is a popular tool for object detection and analysis in many fields. In fisheries, AVI systems designed for fish egg development analysis are used for fish egg assessments in commercial and experimental fish production systems and in ecological monitoring. The first step in AVI systems is to subdivide an image into meaningful regions. The accuracy of this segmentation directly influences the success of subsequent image analysis operations such as feature selection and classification. At present, a lot of image segmentation techniques are available, but there is no universal segmentation technique suitable for all kinds of images.This paper presents a cascading automatic segmentation method for microscopic digitised images of live common carp (Cyprinuscarpio) eggs in Petri dishes filled with water. The proposed method involves five main steps:1) image pre-processing, 2) image initial segmentation by Otsu’s method, 3) segmentation post-processing based on morphological operations, 4) image edge detection and morphological operations to remove incomplete eggs, and finally 5) a Watershed method was used to resolve the problem of separating adjoining fish eggs. Testing the method in Matlab software verified that fish eggs were recognized, segmented and counted with 100% accuracy compared to manual observation in all 96 test cases.

Improving the extraction of cucumber leaf spot disease under complex backgrounds.Redefining the feature distance between pixel xj and clustering center vi.Calculating the two-dimensional neighborhood mean gray value as a sample point.Proposing a new weighting method for gray value and neighborhood gray value. Research reported in this paper aims to improve the extraction of cucumber leaf spot disease under complex backgrounds. An improved fuzzy C-means (FCM) algorithm is proposed in this paper. First, three runs of the marked-watershed algorithm, based on HSI space, are applied to isolate the target leaf. Second, the distance between the pixel xj and the cluster center vi is defined as xj2-vi2. Third, the pixel's neighborhood mean gray value, which constitutes a two-dimensional vector with grayscale information, is calculated as a sample point, rather than FCM grayscale. Finally, the neighborhood mean gray value and pixel gray value are weighted by matrix w. To evaluate the robustness and accuracy of the proposed segmentation method, tests were conducted for 129 cucumber disease images in vegetable disease database. Results show that average segmentation error was only 0.12%. The proposed method provides an effective and robust segmentation means for sorting and grading apples in cucumber disease diagnosis, and it can be easily adapted for other imaging-based agricultural applications.

A new method for extracting key frames of greenhouse videos was proposed.The method took full considerations of characteristics of greenhouse videos.The method combined the visual saliency and online clustering.The IG algorithm was modified by two parameters confirmed by a new measurement. Research reported in this paper aims to improve the identification of greenhouse vegetable diseases based on the greenhouse monitoring video. It presents a method that combines the visual saliency and online clustering to extract the key frame from greenhouse vegetables monitoring video. Firstly X2 histograms are used to measure the similarity of each frame to the first frame, which eliminates the meaningless frames and improve data processing efficiency and costs. Then, all frames will be converted to HSV color space and a saliency map of each frame is generated based on H component value and S component value. According to the saliency map, the salient region can be obtained. During the process of extracting the salient region, there is a possibility that the information of disease spots is lost. Therefore, morphological method would be utilized to restore the lost information. Finally, online clustering is performed to classify the salient regions into different clusters, and mean pixels value is used to select the key frames. The results indicate that this method can obtain information of entire leaf area of vegetables and extract the key frame effectively.

Variable-rate spraying technology can change the application rate by the target information. And the tree canopy volume was a very important target information. An ultrasonic measurement method of canopy volume was presented, and an ultrasonic measurement system of canopy volume was developed. The average of Wheaton and Albrigo manual measurement values was defined as the manual measurement values in this paper.150 peach trees having different canopy volumes were selected to as the experimental objects. Experiment results indicated that the ultrasonic measurement system had high stability (R2=0.968, Std. Error of the Estimate was 0.061719), the manual measurement method had high stability(R2=0.973, Std. Error of the Estimate was 0.060214), and the manual measurement values and the ultrasonic measurement values were in a good agreement(R2=0.968, Std. Error of the Estimate was 0.065314), that is the ultrasonic measurement system had high precision.

This paper presents a method to extract color and texture features of an image quickly for content-based image retrieval (CBIR). First, HSV color space is quantified rationally. Color histogram and texture features based on a co-occurrence matrix are extracted to form feature vectors. Then the characteristics of the global color histogram, local color histogram and texture features are compared and analyzed for CBIR. Based on these works, a CBIR system is designed using color and texture fused features by constructing weights of feature vectors. The relevant retrieval experiments show that the fused features retrieval brings better visual feeling than the single feature retrieval, which means better retrieval results.

Swine-Vet is a kind of a web-based expert system for swine disease diagnosis and has been promoted in north of China past 3years. The few issues relevant with low diagnosis efficiency and error have arisen during the system test and promotion due to the limitation of previous adopted reasoning techniques and immaturity of diagnosis domain knowledge. This paper proposes a hybrid multimodal reasoning mechanism to resolve the swine intercurrent diseases and improve the system diagnosis efficiency. The framework consists of two modules: one is the forward diagnosing module, information fusion method based on evidence theory is used to solve the barriers of syndrome in swine diagnosis field, and the other is the backward confirming module, set pair analysis (SPA) is adopted to resolve evidence conflict and distinguish uncertain and unknown symptom information during the process of diagnosis. Then, in order to demonstrate this model, a case study is given. Finally, the benefits and challenges of the hybrid multimodal reasoning application in expert system are discussed.

Fish disease is becoming the major factor damaging the sustainable development of aquaculture and the quality security of aquatic product in China. Many researchers provide assistances and advice in fish disease diagnosis and appropriate treatment measures. However, the risk of misdiagnosis, incorrectly treatment or over-treatment also exists due to carrying out after disease outbreaks. So it is necessary to develop an early warning system. Taking into account that fish disease outbreak is mainly caused by suboptimal water quality in north of China, the best way is to develop decision aids tool that can assess and early warning the risk of fish disease based on water quality management. This paper introduces a knowledge-based early warning system for fish disease via water quality management. The architecture, the process for knowledge engineering, main components and their functions, especially a series of water quality management mathematic models are described. At present, the system is in pilot in the city of Tianjin in China. The stage achievements in developing the intellectual system of fish disease early warning are summarized.

This paper presents a research attempt to develop and evaluate a Web-based tele-diagnosis system for fish diseases. Fish disease diagnosis is a complicated process and requires high level of expertise. A tele-diagnosis system T-VET is being developed and tested in China. T-VET is the key sub system of a Web-based expert system for fish disease diagnosis. This paper looks at why a tele-diagnosis system is needed and what are the advantages and difficulties in developing and using such a system. The system's architecture and components are described. A user-centered approach was adopted throughout the development process to ensure the system would meet users' needs and will be accepted. As a result, the system functions were emerged and improved by users' requirements. The experience and feedback from testing and evaluating the system with potential users are discussed. Some valuable insights into the use of the Internet as a tele-diagnosis tool in aquaculture are highlighted.

Edge geometric measurement analysis is an important method of image understanding and portraying the target feature. In this paper, we compress 17 interrelated shape descriptors which are based on edge geometric measure into 6 independent components, and discuss their meanings by using principal component analysis. The analyses in this article provide guidance for the shape feature optimization and accurate identification for greenhouse strawberry leaves images successfully.

The system of Case Based Reason is researched in this paper, a kind of genetic algorithm is proposed to solve the problem of feature vector space, and proposes an approach of integration of CBR AND RBR, an example about fish disease diagnose inferring integration model is given.

Dissolved oxygen (DO) concentration plays a very important role in fish life and aquaculture, but DO prediction is very difficult. So a decision support system for DO prediction based on fuzzy model and neural network was attempted. The paper was based on vast monitored data, every day detecting for two years, in aquaculture pond in North China for two years. This is a preliminary attempt towards a wider use of Artificial Neural Networks in the management of aquaculture water quality. It proposes a model to be used effectively in prediction of DO concentration in aquaculture. This is really a crucial task, especially during the long dry summer months. The prediction of potential risk due to low DO is also very important. This data volume was divided in the training subset comprising of 106 cases and in the testing subset containing 26 cases. The input parameters are sunlight, wind speed, temperature, water temperature, air pressure, pH value and NH-NH3. Consequently three structural and seven dynamic factors are considered. After several and extended training-testing efforts a Modular Artificial Neural Network was determined to be the optimal one.

In this paper, genetic algorithms and machine learning theory and method of knowledge acquisition for expert system structure is proposed to solve the "bottleneck" problems in the traditional machine learning methods on the basis of AQ, the model of knowledge acquisition based on GA is proposed and its application is used in the Fish Disease Diagnosis reasoning system, the rules is accessed to fish disease diagnosis. Fish Disease Diagnosis and the problems of knowledge accessed into combinatorial optimization are solved.

A method to express Profile Hidden Markov Model (Profile HMM) parameters with compressing matrix is presented, which is obtained by imposing the characteristics of both the state transfer and the character output in the Profile HMM. Profile HMM, Multiple sequence alignment, Bioinformatics

Fish disease diagnosis is a complicated process and requires high level of expertise, an expert system for fish disease diagnosis is considered as an effective tool to help fish farmers. However, many farmers have no computers and are not able to access the Internet. Telephone and mobile uses increase rapidly, so, the provision of call centre service appears as a sound alternative support channel for farmer to acquire counseling and support. This paper presents a research attempt to develop and evaluate a call center oriented Hybrid disease diagnosis & consulting system (H-Vet) in aquaculture in China. This paper looks at why H-Vet is needed and what are the advantages and difficulties in the developing and using such a system. A machine learning approach is adopted, which helps to acquire knowledge when enhancing expert systems with the user information collected through call center. This paper also proposes a fuzzy Group Support Systems (GSS) framework for acquiring knowledge from individual expert and aggregating knowledge into workgroup knowledge by H-Vet in the situation of difficult disease diagnosis. The system’s architecture and components are described.

There are three kinds of uncertainty in the process of fish-disease diagnosis, such as randomicity, fuzzy and imperfection, which affect the veracity of fish-disease diagnostic conclusion. So, it is important to construct a fish-disease diagnostic model to effectively deal with these uncertainty knowledge's representation and reasoning. In this paper, the well-developed parsimonious covering theory capable of handling randomicity knowledge is extended. A fuzzy inference model capable of handling fuzzy knowledge is proposed, and the corresponding algorithms based the sequence of obtaining manifestations are provided to express imperfection knowledge. In the last, the model is proved to be effective and practicality through a set of fish-disease diagnostic cases.

Neural networks is considered the most powerful classifier and rough set theory is thought of the best to reduce attributes and to generate rules. The combination of neural networks and rough set is very useful for knowledge acquires. Integrating of the advantages of two approaches and having solved the data continuous problem, this paper presents a hybrid method to extract classification rules. There are three models in our method, in first model, neural networks was employed to classify the data sets. In the second model, the continuous attributes are discretized and the self-organizing neural network was applied to ensure result consistent before and after the discretization. In the third model, rough sets theory was used to reduce attributes and generate the rules. The proposed approach was applied on abandoned mine wastes data and the extracted rules was testified based on the analysis of case studies, the result show that the method was reasonable.

This paper presents the comparableness of SVM method to artificial neural networks in the outlier detection problem of high dimensions. Experiments performed on real dataset show that the performance of this method is mostly superior to that of artificial neural networks. The proposed method, SVM served to exemplify that kernel-based learning algorithms can be employed as an efficient method for evaluating the revegetation potentiality of abandoned lands from coal mining activities.