Abstract:Test case selection is an important part in software testing. However, there is little research about test case selection in testing of concurrent programs. We propose an adaptive random testing approach for multi-threaded concurrent programs. By selecting test cases with heuristics, experimental results show that our approach can improve the efficiency and fault detection capability of random testing of concurrent programs.

Abstract:In recent years, the near space hypersonic vehicle is widely concerned by the military in the world. The near space hypersonic vehicle develops quickly. Based on this background, this paper chooses near space hypersonic vehicle as the research object, on the basic principle of maneuvering target tracking based, summarizing the improvement direction of the algorithm of maneuvering target tracking in recent years. Starting from the maneuvering target model and adaptive filtering algorithm, improved algorithm is proposed for the idea, design and development of near space hypersonic velocity of target tracking system to provide support on theory exploration.

Abstract:To ensure the best operation condition of the storage bunker ball mill, we must make its load, temperature, and its inlet negative pressure be closed to the best value. The traditional fuzzy control of storage bunker ball mill is used widely now, but it is very difficult to guarantee the system run in the best condition if the control totally accord to the expert experience. Therefore this passage adopts the genetic algorithm to optimize membership function of the fuzzy control to control the operation of the system. This way can avoid the subjectivity of membership function depending on expert experience.

Abstract:Time overhead is important in real-time operating system. Context switch is one major delay because it occurs frequently in a RTOS. The given DSP has features for having a dual data paths, providing multiple storage blocks can be accessed in parallel, general-purpose register banks can be accessed in parallel. To reduce the overhead of context switching, we proposed a solution based on the features of the given DSP. We split registers of context into two parts, and saved them in two memory parts which can be subjected to parallel access. When context switching, saving and restoring the registers of context via dual data paths is rather than one. This solution had been implemented in the RTEMS operating system. The experimental results show that saving and restoring the registers via dual data paths can reduce the processing time by 49.04% comparing to one data path.

Abstract:Metagenomic binning is a fundamental question for metagenomic studies. Features extraction is the main factor which influences the performance of metagenomic binning, and how to extract the appropriate feature vectors will influence the binning accuracy and running time. Therefore, this paper proposes a features extraction method which based on third-order Markov model and transferring probability matrix for metagenomic binning problem. Meanwhile, we employ the features selection method based on mutual information to reduce the dimensions of feature vectors and apply it to support vector machine algorithm for binning as well as making comparisons among similar binning algorithms. The results show that this new features extraction method possesses applicable discriminability among different metagenomic species, which is particularly appropriate for large-scale metagenomic binning problem.

Abstract:In order to solve the lack of timeliness in barcode recognition's algorithms, the inadequate stability in the application of locating, the insufficient depth of field and etc., it introduce a mixed designed barcode recognition system based on computational vision, FPGA and DSP, the high-speed locating could be achieved with the locating time of only 2.5 ms, reduces the amount of DSP's computation, greatly enhances the timeliness of system and be applied to the tire barcode Recognition system. Experiments results show the system can recognize 16 frames per second with a high recognition rate, which is of practical value in online detection of rapidly moving barcodes.

Abstract:To make the best of the Chinese medical information in electronic medical records, a visual organization method is proposed. Firstly, a medical information dataset based on electronic medical records and medical community web pages is constructed, which is preprocessed into text corpus. Secondly, a topic model using Mallet is trained and visualized the output of topic model. Finally, a visual analysis system for Chinese medical information is also built. Experiments showed that the system could effectively help the analyzers train topic models and diagnose.

Abstract:The paper promotes a cross-platform components integration framework for human-machine interaction(HMI) software and illustrates it from three aspects. First of all, the paper introduces its frame construction which includes bootstrap program, mainframe, component entity and message bus. Then it presents the management approach of all components:each component entity is a cross-platform unit loaded by the mainframe separately with the bootstrap program. The framework derives all component entities dynamically through virtual function system, then mounts them to the message bus with their name as hash keyword and inserts their interfaces into the mainframe interface container. Thirdly the paper elaborates the approach for component status switch and interactive collaboration which is based on the leader-imitator pattern and supports fast message bus and event loop bus. In the approach a component gets focus by mouse or keyboard event, it is selected as active leader and enters dormant imitator status again after processing the event. And all components interact with each other through messages. The framework was already put into application in configuration software for protection and monitoring control device and programming software for industrial control, and was demonstrated observably efficient in reducing coupling relationship of components while boosting the development efficiency and component reusability.

Abstract:Due to the limited network bandwidth and the fact that the 3D animation data are very huge, it will take a long time to transmit the data of 3D animation through network on one time. As a result, it can't be smoothly displayed in real-time as the transmitted 3D animation data are incomplete. The key technology of WebGL based 3D animation is presented, such as model files preload and transmission of compressed files, et al. Under the condition of limited bandwidth, the proposed approach has the capability to transmit 3D animation content to the client in real time. The 3D model is created and rendered by WebGL, and users can view and interact with the 3D digital animation in high-speed through the browser.

Abstract:Embedded software development and debugging tools based on Eclipse platform has important significance for embedded software development, but at present, the number of debugging tool based on Eclipse platform is few. In view of this, this paper provides a new implementation of debugging tool named MDSPTool which supports embedded multi-core system. It is based on Eclipse and plug-in development environment, and can seamlessly integrate with WorkBench IDE. Experimental results show that MDSPTool, which has an excellent user interface, provides easily used methods for parallel debugging and highly integrated functions for monitoring and measurement. Therefore, MDSPTool improves the efficiency of embedded software development.

Abstract:Push notification system plays an important role in the implement of enterprise office automation. And communication protocol and system architecture have become the main factors in the performance optimize of enterprise push notification system. Traditional standardized protocol scheme has encountered problems such as too complex to implement and huge traffic cost in mobile internet. This paper, for the stated problem and the features of enterprise push notification system, designed a communication protocol based on MQTT, a lightweight protocol in Internet of Things, and partitioned social network notification, instant messaging and presence by the design of MQTT topics. The new designed protocol is easy to implement and small traffic cost. On the other hand, this paper, based on the traditional symmetric and asymmetric system architecture, proposed a system architecture which is asymmetric on information and symmetric on function, by which supplied a detailed solution for push notification system for enterprise. This solution can meet the requirements of system design as well as improve the scalability of the system.

Abstract:This paper analyzes the working principle and communication protocol of 8250 serial and driver of 8250 in Linux. Then it builds the camera control system which is based on serial under the video surveillance system on freeway. It completes serial driver design in the Linux, serial application design and application design of control on camera. It implements the control of camera, which is used in video surveillance system on freeway. So, it has good results and high social value.

Abstract:This paper designs and implements the multi-transport mode IP network video monitoring system based on H.264. The ARM9 S3C2440 is used as the system core hardware processor. Video data is captured by the USB camera and encoded by X264. The encoded video data is packed in the network layer by RTP and transmitted in Wired Ethernet, Wi-Fi and Power Line by UDP Socket which is established between ARM-Linux development platform and PC. After receiving the video stream, PC will decode and real-time display. Finally, compression coding efficiency in the server and system performance in three transmission modes are tested and analyzed. The test data and the results show that the system has the advantages of low power consumption, high compression coding efficiency and clear video quality, etc. And it can ensure the high real-time video transmission in various transmission environments.

Abstract:The existing streaming media transmission of remote medical system needs a proprietary network and equipment endorsement. As a result, it cannot be widely applied in real life. This article proposes a remote medical system based on crowd and participatory network which can be applied on traditional network. With the support of the system nodes' topological characteristics in social network, we build the crowd and participatory network model in which the nodes can collaborate unconsciously. Based on this communication model, we design and implement a distributed remote medical system, which don't need proprietary network and equipment. Our method provides a new idea to build remote medical system. The experiments shows that our system can effectively increase the average data transfer rate under the traditional network environment.

Abstract:To improve the traditional manual method of teaching video reservation and administration in university, a teaching video reservation system based on android is proposed to address the requirements of improving the teaching practice ability and classroom teaching effect, improving the work efficiency and saving labor costs. The system is based on the C/S framework and uses PC as a server to provide the underlying service, provides the user-interface for users by the client application in android mobile phones. The client cannot access the database server directly, but introduces the web service interface to the business logic layer. The experiment results show that the system can improve the defect of the traditional appointment method effectively, can improve work efficiency, can bring the convenience to the work among teachers and videotaping staffs, and can enhance information level of education work in colleges and universities greatly.

Abstract:In order to solve the issue of the deterioration of reconstruction performance in compressed sensing(CS) for acceleration signal with poor sparsity that is produced by noise, this paper proposed a new method for perfect reconstruction of acceleration signal by using empirical mode decomposition(EMD) and wavelet denoising. Its basic idea is that the best sparsity of acceleration signal is firstly gained by using EMD and wavelet threshold denoising method. And then, considering CS and acceleration signal with block structure, block sparse Bayesian learning algorithm is applied to perfectly reconstruction original acceleration signal. The acceleration signal from human activity dataset USC-HAD is selected to test the effectiveness of the proposed method. The experimental results show that when compared to traditional CS algorithm without pre-denoising processing, the proposed method can obtain the best value of signal-to-noise ratio and root mean square error. Also, these results suggested that our proposed method has the superior ability of denoise for gaining best sparsity of acceleration signal, which significantly improve the reconstruction performance of CS.

Abstract:General search engines often cause the topic-drift problem, which means that during the retrieval process, some of the retrieval results are independent to the domain keywords. We propose a web page re-ranking algorithm for a specific domain-the TSRR(Topic Sensitive Re-Ranking) algorithm to solve the problem from a specific perspective. TSRR establishes a vector model which is independent to page rank for a specific domain and a web page information model; then it combines the vector model and the web page information model to re-rank the search results in the retrieval process. TSRR's performance is evaluated based on the criteria of customer satisfaction and precision. Experiment results on the dataset crawled for specific domains show that TSRR is excellent in performance. Compared with the ranking algorithm from Lucene, TSRR can promote the customer satisfaction performance by 17.3% and the precision performance by 41.9% on average.

Abstract:Container terminal resources efficiency have been studied for many years, however, most researches focused on offline algorithms which are given the whole problem data from the beginning. Due to some unforeseen factors(such as bad weather, equipment failure), the predict information is not reliable, these make the planned difficult to implement, and reduce the efficiency of port operations. So we are based on the idea of online algorithms and propose a optimization model of berth container terminal and bridge crane resource, and give a online scheduling algorithm for solving. We also analyzed the smoothed competitive ratio by the smooth analysis. Through the experiments confirmed the algorithm feasibility.

Abstract:With the improving of the communication technology and the increasing of the applications, the wireless spectrum resource becomes more and more scarce. The resources utilization rate of the primary users who has been assigned resources is low. The cognitive radio technology hope to improve spectrum utilization in a way that the second users use the the primary users' resource. Spectrum sensing is a key technology of the cognitive radio. This article is based on the existing sensing technology,and put forward a new sensing mean. In cognitive radio network, the respective signal to noise of second users are not same because that their distance to the first user and their own signal environments are different. The dynamic dual-threshold cooperative spectrum sensing method is an improvement over the traditional two-threshold energy-aware approach. It improve the sensing performance by fully considering the difference between the secondary users.It has several steps. First of all, it need to filter out some second users that do not meet the conditions, such as, the probability of detection and false alarm. Then, each second user will dynamic change its dual-threshold which is used for energy detection according to its SNR. Through comparing with the single threshold energy-aware cooperative spectrum sensing method and the traditional dual-threshold cooperative spectrum sensing method, we can find that our method has a better performance on improving the overall performance of the system than both of other two methods.

Abstract:Global gray-level histogram lacks of spatial information, SIFT feature matching method do not have a high accuracy. In view of the mentioned problems, this paper proposed a method of extracting color moment, and advanced the measuring method of SIFT features. Then combine the two advanced method for image retrieval. The experiment result shows that the algorithm proposed in this paper do a good job in image retrieval.

Abstract:To solve the problem that the accure data can't be pushed by the failure of local sensor in intelligence greenhouse system, it was presented that the Apriori algorithm which was based on association rule applied in the prediction of sensor fault data. Forcasting the greenhouse environment temperature is provided as an example in this paper, firstly, the classic Apriori algorithm is modified. Then it was used in the prediction of fault sensor data. The experimental results show that the improved Apriori algorithem could quickly find the association rule between the parameters in Greenhouse, thus estimated the range of parameters of the fault sensor and the method could be proved to be feasible.

Abstract:Compared with traditional monolithic kernel based virtualization architecture, microkernel-based virtualization architecture has smaller trusted computing base and its security is easier to be verified by formal methods completely. However, in microkernel-based virtualization architecture, communication between virtual machines on the same computer still requires the involvement of network card driver, which leads to low efficiency. In this paper, we propose a method to accelerate communication between virtual machines on the same computer by adding data selection module and data transmission module in the network service of microkernel-base virtualization architecture. With the method, communication between virtual machines on same computer can be completed in the memory instead of network card driver. Experiment shows that our method can improve the efficiency of communication effectively.

Abstract:MODV is a general purpose dynamic verification tool for memory consistency model. It implemented a time order based frontier algorithm, which has the lower time complexity. To achieve higher performance of MODV, we design and implement several accelerate strategies for its algorithm. The experimental results show that the performance of enhanced algorithm is better than the baseline algorithm.

Abstract:In order to obtain higher network throughput and lower end-to-end delay, this page designs a MAC protocol which suits cluster network called a receiver-initiated MAC protocol based node state for underwater acoustic sensor networks(RINS), this protocol based on receiver reservation can effectively obtain neighbors' state and dynamically allocate time slot. This protocol can realize data packets arriving at receiver in the way of packet train. The simulation results show that comparing to FERI and RIPT, the RINS can improve network throughput, reduce data loss ratio and end-to-end delay.

Abstract:HDFS provides the underlying storage for Hadoop, however, the HDFS deals with massive small files inefficiently and decreases system performance seriously. To solve this problem, we designed a file merging, indexing and retrieval solution. Then through a series of experiments compared to the original HDFS and HAR solution, it can be shown that our scheme can effectively reduce the memory usage of Namenode and improve the I/O performance of HDFS.

Abstract:Name ambiguity is a common phenomenon when one tries to search for someone's information in the Internet. In this paper, we have studied the web name disambiguation issue in detail. After extracting different features related to the name and then creating combined features by vector space model, we give priority to cluster the documents with high similarity by hierarchical clustering algorithm. Evaluated on the WePS data set, the proposed method showed its effectiveness in solving name disambiguation problem.

Abstract:Function call relations can be used to reveal the dependency relations between functions in software systems. An integrated function call relation makes a good contribution to program verification and deadlock analysis, and improves the completeness of verification and analysis. Existing methods of function call relations based on static analysis do not provide integrated support for function pointers or virtual functions, which lowers the accuracy of analysis result. This paper proposes a static analysis method for function call relations generation based on Control Flow Graph(CFG). This method gains source file information by a GCC plugin, and predicts function analysis paths. Then it uses a simulation algorithm proposed by this paper to analyze the statements of those paths, while the call of function pointers and virtual functions are analyzed accurately. It generates full function call relations based on the analysis result. Experimental results show that this method can support the analysis of function pointers and virtual functions, and improves the accuracy of analysis results.

Abstract:We present a new method for computing invariants for cache coherence protocol in this paper. The verification of cache coherence protocols is always a challenge as we have to check the safety properties of cache coherence protocol with arbitrary nodes. By searching the relations between the invariants and the rules of cache coherence protocol, we can verify the cache coherence protocol with circular reasoning. Besides, we implement a tool to help us computing invariants for cache coherence protocol. We also have experiments on how to apply our method to the German protocol with both control property and data property.

Abstract:The physiological noise such as heart-beating and respiration can cause great interference to the subsequent process of data analysis. Because these noise has considerable autocorrelation characteristics. Combined with the characteristics of physiological noise in time and space, through the method of canonical correlation analysis, the noise components can be steadily identified and removed from the nerve tissues in the residual data after regressing the useful signal components. In this paper, the proposed method does not require any priori information about experiments, realizing the unsupervised reduction of fMRI physiological noise. Through the analysis on the real fMRI data, we illustrate the effectiveness and the reliability of our method.

Abstract:This paper proposes a rapid image matting method based on categorized video sequences. The image sequences in a video are classified into three categories:key frames, trivial-changed frames and variant-changed frames. Different image matting algorithms are adopted to treat each image sequences category. First, the method uses closed-from image matting algorithm to process the key frames and get the alphas, foreground colors and background colors. Then, the proposed method uses the known values of key frames to estimate and optimize the alphas of the trivial-frames based on frame-coherence. After that, the matting algorithm is processed by utilizing a SIFT Flow based information propagation strategy to transfer informs from key frames to variant-changed frames. The proposed method achieves acceptable matting results and more higher efficiency compares to traditional methods.

Abstract:Single-source shortest path problem is one of basic researches in graphic theory. Considering the similarity and easy transferability of data structure between graphic and tree, an improved algorithm solving this problem is put forward based on the technology of XML. Using XML structure, this algorithm transfers a graphic to some tree, and analyzes each edge in this tree to adjust the best distance between other vertices and the source vertex. With basic controls used in the execution of this algorithm to simulated the process, and experimental results show that this new algorithm is much more efficient, and more flexible, and easy to understand and good to practical applications.

Abstract:Congestion phenomenon occurs frequently in wireless monitoring network as excessive load of some local node, which is caused by some negative factors such as attenuation, multipath, blind and so on, and then lead to reliable data transmission can not be guaranteed. From the view of congestion control, a predictable based congestion control algorithms based on the design idea of initiative to avoid the congestion is proposed to realize adaptive congestion control which considers link quality indicator LQI when selecting a new route, and obtains the network traffic forecast future values according time series ARMA prediction algorithm, thus determine the congestion degree of a node to take measures in advance. The experimental results show that, the algorithm can improve the reliability of the wireless monitoring system of data transmission, and can improve the throughput.

Abstract:In this paper, the weight of the incremental mining thinking optimization algorithm, for users to recommend personalized product configuration provides an effective solution. The method is mainly divided into three parts, first using the platform to build up user tracking module for tracking user behavior and collecting data; then combined with the user's behavior recently, the use of association rules mining based on the weight increment Apriori algorithm; final complete the product according to the recommended procedure to dig out results. By mining algorithm optimization, greatly improving the efficiency and accuracy of the product is recommended with the change in user behavior and changes in the system, more in line with the actual situation. Experimental results show that the algorithm can effectively solve the problem of product recommendation, compared to the traditional association rule mining algorithm, the accuracy is improved by 4%.

Abstract:Confronting the situation of uneven educational background, knowledge comprehension and master speed of soldiers, stereotype education system of soldier's occupational skill no longer adapts the demand of network era development and informational military construction. The paper puts forward to the tactic of dynamically call ant algorithm and genetic algorithm according to the best fusion point after analyzing the existing problems in soldier's occupational skill education and the features of ant algorithm and genetic algorithm, so as to urge the optimization results of ant algorithm to optimize the initializing population of genetic algorithm. In addition, in order to find the personalized learning path suited to every soldier, the personalized learning process of soldiers is transformed into a typical combinatorial optimization problem successfully by imitating traveling salesman problem. The experiment results show that the convergence rate and optimization capability of the improved ant colony genetic algorithm is greatly improved.

Abstract:To monitor the operation of embedded devices in harsh outdoor environment, this thesis designed an embedded E-mail advice software based on SMTP protocol. E-mail notification software is running on embedded devices with GPRS network, it first goes for the test to get the warning information and forms the E-mail's content and title in designed format, then establishes TCP link between the device and E-mail proxy Server through the GPRS network and sends the warning information to the E-mail proxy server according to the grammatical pattern of the SMTP. At last it designes a relative experiment to confirm the reliability of E-mail notification software. Tthe result of experiment shows that embedded E-mail advice software based on SMTP can meet the real operation's requirement and it has a relative reliability and stability.

Abstract:In this paper, a new algorithm of taillight signals detection and recognition based on image is proposed. Firstly, the algorithm detects and tracks vehicle taillights using color and symmetry features of the images. Secondly, we employ parameters-optimized Least Squares Support Vector Machines(LS-SVM) to classify the current states of detected taillights. Finally, we set up state chains from past frames to recognize the signals of taillights. Through the experiments on the real road, we can see that the recognized signals can correctly indicate the behaviors of brake and turn of preceding cars, which shows that our algorithm is valid.

Abstract:The fuzzy C-means clustering algorithm is the most widely used clustering algorithm, but it still remains sensitive to outliers and dependent on initial centers and other issues. Therefore, this paper presents an improving fuzzy clustering algorithm based on sample weighting, the algorithm can get more accurate initial center points and remove noise. At the same time, to the weakness of the clustering algorithm in Weka system and the clustering problem is extensive in the field of data mining, this paper makes the platform the secondary development, researches the traditional FCM algorithm and improving algorithm. The study finds that the improving algorithm makes the clustering results stable, obtain the accurate clustering results and improve the accuracy of the algorithm.

Abstract:Gradient vector flow has achieved good results in image processing. But the simple iterative computing method, result in slow convergence speed, restrictions its applications. In order to explore the effective solving scheme for the gradient vector field, this paper proposes a method based on BFGS algorithm for gradient vector field calculation. The numerical solving process is given in detail and which is used in image processing task. Experimental results show that the BFGS-GVF algorithm shows a better performance of gradient vector field. Compared with generlized Newton-Type methods for energy formulation, CV active contour, IALM-GVF Snake algorithm, BFGS-GVF Snake algorithm can obtain clear and smooth contour.

Abstract:Decision tree is a very popular method of data mining. This paper introduces the theory of decision tree to analyze the structure of the decision tree, C5.0 algorithm discussed the idea of their advantages and disadvantages. in-depth understanding of the main factors of psychological symptoms and mental health of college students, the C5.0 algorithm is applied mental Health assessment data mining results based on a better understanding of students' mental health problems, to carry out a mental health education colleges and universities have practical significance.

Abstract:An increased adoption of cloud computing has lead to a greater concentration of hardware in massive named datacenters.It is essential for these datacenters to be energy efficient not only to cut down on electricity costs but also to be in compliance with environmental regulations.The author implemented an enhanced version of Hadoop YARN framework that utilizes RAPL's power capping feature to mitigate computational imbalances in an application and to reduce CPU power consumption, named Phadoop. The purpose of the experiment is investgating whether it is beneficial to use RAPL interfaces to conserve the energy consumption of a CPU in a cloud-based workload without significant loss of performance. Experimental results indicate a reduction in energy consumption of Phadoop up to 34% compared to Hadoop.

Abstract:Intelligent tutoring is a hot field in the study of information-based teaching. Applying this technology to knowledge of regulations and operational procedures that are expected to remembered accurately and obeyed strictly, will improve the ability of grasping and using related knowledge. Using the method of object oriented, improve greatly the flexibility of general theme representation and question template. Template-based methods of domain knowledge representation and tutoring questions generation are all presented. It can generate multiple tutoring questions for different attributes. The reduced state transition method can avoid efficiently the problem of state explosion. Finally, taking shipboard damage control as one application background, a prototype system was developed and worked well using this method, illustrating the feasibility and practicality of the design method.

Abstract:Quadrotor UAV is a research hotspot in recent years, who is widely used in many fields because of its advantages in simple structure and vertical take-off and landing. The control effect of classical control algorithm, such as PID Control algorithm, is not satisfying when the UAV is under disturbance or in a changing condition, the UAV can't be adjusted to equilibrium state when it's under large disturbance. This paper modeled through dynamics and properly simplified it, designed the adaptive control law, and the numerical experiment shows that model reference adaptive control can block out disturbance, verifies that this algorithm is effective.

Abstract:On the issue of the target object location in three-dimensional space, this paper proposes a method to realize binocular visual positioning with monocular, this method is simple, convenient operation and cost-effective. Because SURF algorithm has better performance and indicators in all respects, so we make use of SURF algorithm to realize the image feature point extraction and matching in the process of object target recognition and locating. The experimental result indicates, the method of monocular to realise binocular visual positioning based on SURF algorithm, both in accuracy of positioning, and in speed of time shows good feasibility and practicability, it has a certain practical value.

Abstract:Extreme learning machine(ELM) is a new type of single feed layer neural network algorithm. In the training process ELM only needs to set the hidden layer node number of suitable, random set the input weights and hidden layer deviation, finish in one time without iteration. Now use the genetic algorithm to optimize the extreme learning machine to find the optimal parameter values, so as to establish the Chengdu Shuangliu International Airport passenger throughput prediction model. Then through the comparison of support vector machine, BP neural network, analysis the feasibility and advantage of genetic-extreme learning machine algorithm. The simulation results show that the genetic-extreme learning machine algorithm is not only feasible, and compared with the original extreme learning machine algorithm, it has obvious advantages in prediction accuracy and training speed.

Abstract:With the influence of shadow on plant leaf images in natural environment, extraction of foliar features will be seriously affected. Hence, this paper proposes a method of shadow detection and removal based on YCbCr color space. First, the intensity of channel Y in YCbCr color space is calculated and the shadow regions are detected by a threshold on the intensity. Then according to the shadow model, light restorations for each pixel in shadow regions are made and the results are finally transformed to the RGB color space. Compared with direct shadow removal in the RGB space, this method weakens the edge effect, thus making the color between the post removal shadow region and non shadow region more uniform and the restored images more natural.

Abstract:This article describes the composition and implementation of the Report Designer flex technology-based. Firstly it described the use and advantages of flex technology industry, then it introduced the common problem in the industry. It described in detail the various modules based report designer and flex technology the corresponding function, and by this designer to design the whole process from the report on the report line. Finally elaborated this designer next direction for improvement.

Abstract:Data security is an important problem in digital campus construction. It is urgent to find a way to detect the risk and vulnerabilities of digital campus security. Therefore, a digital campus etwork security detection prototype was designed based on the improved BP algorithm. The data flow and nformation coming from the underlying network protocols(TCP) will be collected and pre-handled. The trained neural network module will react to these data and information, which gives the clue to determine the existing attack or scanning behavior in current network data flow. Rapid detection of digital campus existing vulnerabilities and security risks can prevent and reduce the attack and destruction.