Sample collection method and device based on distributed system, equipment and medium

Abstract

The embodiment of the invention discloses a sample collection method and device based on a distributed system, equipment and a medium, and relates to the technical field of data collection. The methodcomprises the following steps: receiving a sample collection instruction, and setting a timing collection task; acquiring equipment information of the collection equipment and synchronizing the equipment information to locally stored equipment information of historical collection equipment; triggering a timing collection task; reading device information of the target collection device to obtain adevice identifier; and identifying the target collection equipment and establishing communication connection, accessing the target collection equipment to obtain a video stream of the target collection equipment, and capturing a plurality of pieces of sample data from the video stream for storage. The invention also relates to a blockchain technology, and the sample data is stored in the blockchain network. According to the method, through reasonable time period arrangement in a real scene and timing starting of distributed task snapshot scene sample data, automatic sample collection is realized, the operation amount of workers is reduced, and the speed and quality of sample collection are greatly improved.

Description

technical field [0001] The present application relates to the technical field of data collection, in particular to a sample collection method, device, equipment and medium based on a distributed system. Background technique [0002] Artificial intelligence has three elements, which are data, computing power and algorithms. Because the foundation of artificial intelligence is training, only through a lot of training can the neural network summarize the rules and apply them to new samples. Specific to the field of target recognition, the data is image samples. For target recognition that requires deep learning, the larger the network, the more layers, and the more training samples required. [0003] The traditional methods of collecting samples mainly include the following: 1. Obtain from public data sets; 2. Obtain from image libraries on the Internet, such as Baidu Pictures; 3. Purchase from data providers. [0004] For the first and second methods above, the number of re...

AI Technical Summary

Problems solved by technology

[0004] For the first and second methods above, the number of real scene samples that can be collected is limited, and the speed of manual collection alone is too slow

In addition, the quality of the samples obtained by these two methods is not high, whether it is the resolution or size of the picture, it is quite different from the monitoring picture in the real scene

As for the third method, there is currently no relevant data provider in the market that can provide samples that meet some specific real scenarios, such as the samples required by the kitchen video intelligent recognition system, which is used to meet the training of the kitchen violation scene recognition model

Images