Research And Implementation Of Tomato Leaf Pest Retrieval System Based On Hash Learnin

Tomato as a globally important cash crop is prone to pests and diseases during its natural growth,and the pests and diseases can directly affect the quality and yield of the crop,thus bringing huge economic losses to farmers.Although the pest retrieval methods based on hash learning,image processing and image retrieval techniques have shown high practical value and good application prospects in many tasks,there are still some problems.For example,in image noise retrieval,most of the existing research is based on noiseless pest images,and the retrieval research for noisy images is not perfect;in pest identification,the existing retrieval methods tend to ignore their originally similar feature information,and the retrieval effect is not good;in hash learning,the existing crop pest image retrieval technology research fails to effectively combine hash learning,and the method is still in its infancy in the field of crop pest image The method is still in its initial stage in the field of crop pest image retrieval.This thesis focuses on noisy crop pest and disease images to address the above problems.Firstly,the noisy pest and disease images are mapped into Hamming space,and then the Hamming distance is used for image retrieval,and finally the identification and diagnosis of pests and diseases are realized.The main research contents and innovation points of the thesis are as follows:(1)The image of pests and diseases is greatly affected by many factors(such as natural light,transmission and communication),and there are many problems when extracting features directly from the original image,such as large amount of calculation,feature redundancy,large amount of storage,etc.First of all,this paper establishes a pest image retrieval method based on hash learning,which can significantly reduce the data storage and communication costs,thus effectively improving the efficiency of the learning system;Secondly,this paper introduces the Cauchy loss function with smaller reconstruction error to design a robust hash algorithm to achieve better image retrieval for noisy images.(2)Because the traditional hash learning method often falls into the local optimal solution in the optimization process,which will greatly reduce the retrieval effect.Therefore,the swarm intelligence method is introduced into the proposed hash learning method,so that it can intelligently jump out of the local optimal solution and reach the global optimal value to achieve better retrieval results.(3)Based on the above research,a hash learning based image retrieval system for tomato leaf pests and diseases is designed and implemented.The system is mainly to provide a visual research platform for hash learning in crop pest and disease image retrieval algorithm research.The platform not only facilitates scholars to study image retrieval of tomato leaf pests and diseases based on hash learning,but also provides "image search" service for general users.