Tobacco leaf mildew detection system based on block chain and deep neural network

Abstract

The invention discloses a tobacco leaf mildew detection system based on a block chain and a deep neural network. The system comprises a rough semantic segmentation unit, a mildewing part image acquisition unit, a mildewing judgment unit, a mildewing position acquisition unit and a mildewing result display unit, all nodes in a cloud server cluster load parameters and weights required by a mildewingdetection network, and a mildewing detection network reasoning block private chain is generated; for each neural network reasoning request, available nodes are randomly selected from the cloud servercluster, a node chain sequence matched with the mildew detection network reasoning block private chain is obtained, mildew detection network reasoning is executed in parallel, and therefore tobacco mildew detection is achieved. By means of the method, in moldy tobacco leaf detection, the detection efficiency and the detection precision can be improved, and meanwhile the safety performance in thedata processing process is improved.

Description

technical field [0001] The invention relates to the technical fields of artificial intelligence and block chain, in particular to a tobacco mildew detection system based on block chain and deep neural network. Background technique [0002] my country is a big country in tobacco leaf production and consumption. In the production of cigarettes, the tobacco leaf raw materials must first be selected and classified. The quality of the selected tobacco leaves is directly related to the purity and quality of the produced cigarettes. Therefore, it is very important to select and classify tobacco leaf raw materials. Whether the tobacco leaves are moldy or not is a key indicator for classifying tobacco leaves. At present, there are researches on mildew detection based on image processing technology, but the detection accuracy and detection efficiency are often not high. Moreover, in the mildew detection process, factors such as security are not considered, and the processed data i...

AI Technical Summary

Problems solved by technology

At present, there are researches on mildew detection based on image processing technology, but the detection accuracy and detection efficiency are often not high

Moreover, in the process of mildew detection, factors such as security are not considered, and the processed data is easily attacked and intercepted

[0003] Therefore, the existing tobacco mildew detection technology has the problems of low detection accuracy and detection efficiency, and low security in the data processing process.

Images