Study On Wheat Seed Selection System Based On Image Recognition Technology

With the progress of society,agriculture,as the first industry,is also facing technical breakthroughs and new technology application problems.How to realize the automation of wheat seed selection has become the main research object of this paper.Through the research of wheat seed selection method,it is found that the main classification method of wheat seed selection is the body characteristics,color characteristics and quality characteristics.With the help of reading relevant materials and documents,consulting relevant experts and investigating relevant producers,it is found that image recognition technology has become a feasible scheme to identify and grasp the shape,color and quality characteristics of wheat seeds.Opencv,as an opensource computer vision library,has been selected as the main tool of wheat seed feature recognition by the author.On the hardware equipment of wheat seed image acquisition,after many comparisons and combined with the actual application,the author chose the high-resolution instrument which has been mature in the market.On the hardware equipment of wheat seed grading,the author chooses a three-axis mechanical slide controller based on Modbus protocol for serial communication.The three-axis mechanical slide can move in three dimensions.In the process of the experiment,the author first completed the compilation of the software,including the capture of the wheat seed image,the extraction of the wheat seed features,the establishment of the specimen database and the three-dimensional motion control instructions of the three-axis controller,which were encapsulated as a dynamic database.Then,the author selects the known grades of wheat seeds for feature extraction,and establishes the corresponding grade standards according to the shape features,color features and quality features.In the process of wheat seed recognition,through the comparison with this standard,the classification of wheat seeds is completed.Through the control of three-axis controller,the classification of wheat solid seeds is completed.In the process of testing,100 wheat seeds of three grades of good and bad were selected for testing,and the recognition rate of excellent seeds was 97%,good seeds was 87%,poor seeds was 95%,and the comprehensive recognition accuracy was 93%.Through the test data,we can confirm that it is feasible to use the combination of software and hardware to complete the operation of image acquisition,feature extraction,data comparison,entity classification of wheat seeds,and finally realize the automation of wheat seed classification.