Focused on the lighting effect and lighting efficiency in current VLC ( Visible Light Communication) system，a VLC system with QDs-WLED ( Quantum Dots White LED) as light source is proposed，and a pair of portable VLC transmitter and receiver are designed and realized． The core-shell CdSe quantum dots with high stability and high quantum efficiency at 570 nm are synthesized by one-step synthesis method，and its QY ( Quantum Yield) can reach 81%． The quantum dots material，which replaces the traditional phosphor material，is mixed with epoxy resin and then apply to the blue LED ( Light Emiting Diode) chip to make white LED device． We test its emission spectrum，color coordinate map and lumen efficiency． The VLC structure and principle are described，the communication software and protocol are compiled，and both hardware and software are integrated． With the developed devices，the data transmission experiment is performed ． Experimental result show that the quantum dots white LED possess excellent lighting effects and energy-saving features，and also can achieve the function of data transmission．Under the optimum DC bias voltage of 2．70 V，the maximum communication distance of the system is about 1．3 m，the maximum bit rate is up to 267 kbit /s，and BEＲ( Bit Error Ｒatio) is less than 10 ^－3 ．

Coastal ship communication faces the difficulty of medium absorption of high-frequency electromagnetic waves( HF) ，scattering of electromagnetic waves by sea surface turbulence，and multipath signal interference caused by reflection of electromagnetic waves by sea objects． According to the characteristics of the shore-ship communication communication channel，the channel transmission model is established，and the unit impulse response of the channel is solved． The method of the receiver based on double-angler anti-multipath interference in the ship-to-ship communication is proposed． By utilizing the energy of the multipath signal，the communication quality can be improved without increasing the transmitter power，and the anti-multipath effect is good and easy to implement． The high-frequency electromagnetic wave ionospheric reflection model is established． The solution triangle communication algorithm is used to study the time that shore-to-ship communication can maintain communication，which is convenient for improving communication protocol and improving communication quality． The simulation results show that the longest day can maintain 4． 5 h of communication，and the longest can maintain 7． 5 h of communication at night．

Memristive neural networks can effectively improve the complexity of traditional neural network circuits，the difficulty of integration and high energy consumption． Membrane，memristive neural networks and the application of current memristive neural networks in image processing are summarized． Based on the memristive property，the dynamic variable of the neural network synapse is realized，which makes the memristive neural networks having more advantages and wider application range than the traditional neural networks in the field of image processing． The future development prospects of memristive neural networks are forecasted．

Aiming at the problems of potential gas leakage and dust explosion in coal mine underground pipelines，we proposed a pipeline safety monitoring method based on Ｒ-OTDＲ ( Ｒaman Optical Time DomainＲeflectometry) distributed optical fiber temperature sensor，and analyzed the technical principle of Ｒ-OTDＲ fiber． For the problem of temperature signal noise interference during the acquisition process，the two methods of moving average filtering and EEMD( Ensemble Empirical Mode Decomposition) are used for denoising． The simulation results show that the best noise reduction effect can not be obtained by using one method alone． This paper will combine the sliding average filtering and EEMD denoising algorithm，and use three indicators for effect determination． The experimental results show that the first moving average EEMD can effectively remove the interference noise in the signal，which greatly improves the signal-to-noise ratio and system measurement accuracy． Compared to the traditional algorithm，the system sensing efficiency is higher．

In order to realize the detection and control of carbon dioxide ( CO2) in horticulture，a CO2 measurement and control system based on Na + superionic conductor ( NASICON) is developed，which includes CO2 measurement nodes，a wireless communication network，and a control terminal． The obtained CO2 concentration signals are conducted by the STM32 central processing unit and the release of CO2 is controlled by fuzzy PID ( Proportional Integration Differential) algorithm． As a result the CO2 concentration was regulated intelligently to reach the required value for crop． The sensor nodes are calibrated and stability test is performed in laboratory conditions． The relative fluctuation is less than 2． 5% for the detection of 400 × 10 － 6 and 1 000 × 10 － 6 CO2 over a time period of more than 8 h． A field test of the CO2 measurement and control system is performed in Modern Agriculture Greenhouse，Guoxin Group，Changchun，Jilin province． With concentration control，the fluctuation range is about ± 40 × 10 － 6 when the expected value of CO2 concentration is 800 × 10 － 6 ． The field test indicates that the measurement and control system possesses the advantages of high practicability，low cost and compact size，which has a certain value for generalization and application．

In order to reduce the influence of the high linear correlation of each color component in the ＲGB( Ｒed Green Blue) space and the scale correlation of the Euclidean distance to the image segmentation results in classical k-means algorithm，the Lab color space can overcome the defect of uneven color distribution in ＲGB space，an improved k-means clustering method for color image segmentation based on Lab color space is proposed． Firstly，the color space is converted from ＲGB to Lab，and each pixel can be represented by the combination of L，a and b． Secondly，mahalanobis distance is used to replace Euclidean distance，and the improved k-means algorithm is used to cluster the pixels of the image to achieve the purpose of segmentation．Finally，experiment results show that the improved algorithm has better segmentation effect and accuracy than the classical k-means algorithm．

In order to solve the problem that vehicles may hit the coming object on the road around a mountain，we put forward the system for mountain roads anti-collision and monitoring in targeted poverty alleviation area based on LBS． Android software is used to simulation the Internet + terminal for pedestrians and vehicles and Baidu Trace service is used as the main background for testing the availability of the system． This system is suitable for anti-collision monitoring on the winding mountain road traffic in targeted poverty alleviation area． It can calculate the location of dynamic vehicle or pedestrians that are out of the anti-collision safe distance and realize a global visualization so that pedestrians and vehicles in the narrow winding road can prepare for whistle or avoiding in advance relying on various moving trajectory and motion trial data forms from the background of Baidu Trace and make analyses for traffic monitoring and warning． The experiment result shows that the system is effective． Its alarm can reduce the probability of collision on the road around a mountain．

LDAG( DAG algorithm based on linear threshold) algorithm is a heuristic algorithm for maximizing the influence of social networks． It has the characteristics of high accuracy and high efficiency． When solving the problem of maximizing the influence of social networks，the network influence propagation model is given priority，and then the topological structure of social networks is ignored． In this paper，the LDAG algorithm is improved by using the structure of social network community to select the key nodes of influence propagation．The key nodes are used to simplify the construction process of directed acyclic graph，optimize the time complexity and space complexity of the algorithm，and validate the rationality of the algorithm by using two effective experimental data sets．

In order to solve the real-time problem in the process of vehicle identification ahead，a vehicle identification method based on license plate detection is proposed． The method is divided into three steps． First，the region of interest is extracted using details such as a road surface or a lane line in the image． Secondly，HSV( Hue-Saturation-Value) color space conversion and rectangular image detection are used to filter illumination changes，shadows and cluttered backgrounds from the extracted regions of interest，and vehicle license plate information is detected． The secondary positioning is performed and verified in the case where the initial detection fails． Finally，the detected vehicle information is used to identify the vehicle in front． The proposed method is evaluated on self-built and public database videos． The experimental results prove the effectiveness of the proposed forward vehicle identification method， with over 90% recognition rate and high real-time performance．

In order to make the traditional game of Go more convenient and fast，we improve the“influence model”algorithm，and design a kind of row-scanning circuit type intelligent electronic chessboard with windows operating system． Its composition includes three parts: the row and column scanning board circuit，the information acquisition module and the upper computer ( PC) software． The game board information is obtained through the row and column scanning board circuit，and the information is transmitted to the host computer through the information acquisition module，and compiled by the VC program in the host computer to complete the game board display，the player＇s time record，the mesh count，the chess piece count，the outcome Judging and other functions． Experiment shows that the system can stably display the game of chess，and can quickly and effectively determine the outcome of the game．

An improved collaborative filtering algorithm is proposed to solve the problem that the accuracy of recommended results decreases when the sparse degree of score matrix becomes larger in the alternating least square method． According to the user's rating of various Chaozhou delicious foods，the algorithm combines the interest similarity of other users，and uses the similarity of Chaozhou food attributes as the weight factor to complete the matrix． The experimental results show that the average MAE( Mean Absolute Error) value of the improved algorithm is 0． 583，which effectively improves the recommendation accuracy．

In order to solve the problem of classification and accuracy of epilepsy EEG( Electroencephalogram)signals，frequency slice wavelet transform was used to reconstruct EEG data and get five frequency bands of rhythmic signals． We use approximate entropy of non-linear index and fluctuation index of linear index as the eigenvalues of epileptic signals to fully extract the characteristic information of signals． Gradient lifting tree algorithm was used to classify the feature data set． The classification recognition rate of epileptic EEG signals is 98． 4%． Compared to the traditional Adaboost algorithm，we adopt GBDT( Gradient Boosting Decision Tree) as a classification algorithm． This method can use more data sets successfully and has higher classification accuracy．

To solve the problem that optical measurement equipment cannot carry out practical training due to lack of cooperative target close to the actual situation of weapon equipment test，a practical training method for optical measurement equipment based on simulated target is proposed． According to the actual situation of optical measurement equipment and target，the three-dimensional model of the target is constructed with the visible light
and infrared imaging characteristics processing on the three-dimensional model respectively． The third-order
B-spline curve is used to fit the trajectory curve of the target，and the attitude control information is obtained after
the motion attitude processing． The motion characteristic model of the target is generated by encapsulating the
trajectory curve with the attitude control and characteristic events information． Using computer simulation
technology，a virtual three-dimensional space is constructed，in which the scene of optical measurement
equipment tracking and measuring target is simulated，and the simulation target is seamlessly generated on
optical measurement equipment for practical training of optical measurement equipment． The application and
verification on the in-service optical measurement equipment show that the method is feasible and practical．

In order to monitor ECG ( Electro Cardio Gram) signal of sub-healthy people in a real-time，the research used the integrated chip of ads1191 to collect two heart signals ( ＲA and LA) ，which was transmitted to the MSP430F2132 via SPI ( Serial Peripheral Interface) and then to the computer by bluetooth technology． The required heart rate was obtained by using low-pass filter to eliminate interference signal，and using the second order difference method to monitor the position of Ｒ wave in the ECG waveform． Instead of using lead wire between Electrode patch andelectrocardiograph，this research avoids interference caused by it effectively，and meets the needs of daily wear for its smaller size and lower power． Under the battery power supply of 1 000 mAh，the system can work continuously for 48 h，ensuring the normal data transmission within the range of 3 ～ 4 m．

Aiming at the current situation that floating garbage can not be cleaned efficiently and rapidly，this article studies on the control system of the small-scale float-garbage automatic cruise ship based on the open-water traversal algorithm，and an open-water traversal algorithm is proposed which fuses the magnetometer position error compensation algorithm， the GPS ( Global Position System ) azimuth， distance algorithm and the Boustrophedon algorithm，while introduces the close-cycle control to optimize the performance of the control system of the float-garbage cruise ship． The system realizes the acquisition of the position and the system pose information through the GPS sensor and the electronic compass sensor，and the route of the float-garbage cruise ship is controtted after the processing with the open-water traversal algorithm to realize the autonomous navigation． The use of the PID( Proportion Integral Differential) control algorithm in aim of controlling the ship propulsion actuator effectively reduces the occurrence of the border-crossing of the margin． Simulation experiments all the effectiveness of the traversal algorithms has been proved theoretically，the physical tests based on this shows that the traversal coverage ratio of this system in a 63-square-meter polygon water region reaches 95%，and the system could realize the efficient clearing of small-scale water floating materials，has a broad application prospective in the float-garbage clearing field．

In order to solve the lack of an integrated control system with complete functions and high reliability in China's underground pipeline gallery at the present stage，the emergency system is weak，and the safety and reliability are poor． We propose an integrated control system for underground pipe corridors based on CAN ( Controller Area Network) Bus comunication． After several simulation experiments and test platform tests，this system is more reliable and convenient than the integrated control system applied to underground pipe corridors in China at this stage． Extended，lower cost，etc．