Research On Remote Monitoring System Of Plug Seedling Quality Based On Machine Vision

With the increasing demand for automation in modern agriculture,machine vision has been used in agricultural production instead of manual detection and recognition.A remote monitoring system for tomato seedling quality based on machine vision was studied,which can be used to detect the quality of tomato seedlings and to facilitate managers to obtain the results remotely.The main contents of this paper include the following aspects:Firstly,according to the design requirements,carried out the overall design of the system,which can divided into seedling detection subsystem based on machine vision and remote monitoring subsystem based on B/S mode,and designed the composition of the subsystems.Secondly,aim at the layout of LED in machine vision system,build mathematical model with the objective function of minimizing the variance of illumination between discrete areas in the working face.Use genetic algorithm to optimize the calculation.The results show that the measured illumination distribution is similar to the calculated one,and the layout parameters obtained by this method can meet requirements.Thirdly,studied the method of seedling quality detection using machine vision technolog.The image processing technology was used to check the unexplored holes and to classify the seedling quality.Using the top-view image of tomato seedlings in three days,the seedling emergence detecd by extracting the leaf area characteristics of seedlings in each hole.The experimental results show that the accuracy of the algorithm is 97.2%.For individual seedlings,the leaf area,plant height and ground diameter characteristics of seedlings were extracted from their face and top views,use the random forest classification algorithm classtified the quality of seedlings.The results show that the overall accuracy of the algorithm is 93.3%.Finally,the monitoring system based on B/S mode is designed.Using Django framework to design Web application,completed the function of remote acquisition of seedling detection results by browser client.