Image recognition method for agricultural trade product types

Abstract

The invention discloses an image recognition method for agricultural trade product types. The method comprises the following steps of S1, obtaining an image data set of the agricultural trade products; s2, dividing the image data set into a training set and a test set, performing data preprocessing and enhancement on the images in the data set, and then performing size unification on the images; S3, sending the image data set processed in the step S2 into a deep convolutional neural network for training; S4, when the network loss function converges, finding some corresponding types of images in the auxiliary training set according to the types of the wrong images, adding the images into the training set, and continuing to train the network; S5, repeating the step S4 until the network lossfunction converges or the detection result on the verification set is stable; and S6, testing the test set by the trained network model, and outputting a result to judge the type to which the agricultural trade product belongs. According to the present invention, the agricultural trade product type can be quickly and accurately identified, the intelligence of the agricultural trade market can be quickly promoted, and a large amount of manpower and material resources are saved for enterprises.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to an image recognition method for types of agricultural products, which can be applied to intelligent electronic scales. Background technique [0002] With the development of science and technology, more and more smart devices play a very important role in our lives, such as face attendance to replace manual attendance and save a lot of manpower and material resources. When we are shopping in the supermarket, we often find that there are queues to weigh items in the fruit area, vegetable area and other areas. For the intelligent age, we can design an intelligent electronic weighing device to solve this situation. Use smart electronic scale equipment to replace manual identification of the types of agricultural products, and calculate the price according to the unit price of each agricultural product and the weight of agricultural products in the system. Using this electro...

AI Technical Summary

Problems solved by technology

[0006] 1) Manually identifying the types of agricultural products consumes a lot of manpower, material resources and financial resources;

[0007] 2) Workers are prone to mistakenly identify the types of agricultural products when they are tired;

[0008] 3) The barcode may be wrongly pasted manually, resulting in a discrepancy in the price

[0012] 1) The low-level visual features used for the color features, shape features, and texture features of agricultural products, and how to choose a classifier are all based on experience, and it is difficult to guarantee the best detection effect;

[0013] 2) The recognition calculation is time-consuming and not efficient;

[0014] 3) The method of using specific features in advance and then sending them to the classifier can generally only be applied in specific scenarios, with poor robustness and low practical value;

[0015] 4) This method will be cumbersome for the actual implementation site, resulting in an increase in application costs;

[0016] 5) The actual application environment is complex and the light is different, and the recognition accuracy of agricultural products is very low

Images