Device and method for detecting image recognition capability of intelligent monitoring and photographing terminal

Abstract

The invention discloses a device and method for detecting the image recognition capability of an intelligent monitoring and photographing terminal, and the device comprises a hardware part and a software part, and is characterized in that the hardware part comprises a power module, an industrial personal computer, an adjustable power supply, an eight-path current monitor, a switch, and a database / sample library management platform; the power supply module is electrically connected with the industrial personal computer and the adjustable power supply, inputs 220V AC commercial power, and supplies power to the industrial personal computer, the adjustable power supply and the eight-path current monitor after AC-DC conversion; the software part comprises a to-be-tested terminal management function module, a system basic management function module and a sample data management function module. According to the method, the test sample set is flexibly constructed according to the test requirements, the algorithm model and the terminal to be tested can be subjected to integrated evaluation, and the test efficiency and the automation level are effectively improved through the convenient test process and the multi-path test design.

Description

technical field [0001] The invention relates to a device and method for detecting the image recognition ability of an intelligent monitoring terminal, and belongs to the technical field of electric power artificial intelligence model evaluation. Background technique [0002] With the vigorous development of artificial intelligence technology, the demand for image recognition and other technologies in power inspection, safety supervision, marketing and other professions is becoming stronger and stronger. It is urgent to apply the advanced research results of intelligent image recognition technology to the power field to achieve cost reduction and increase. effective goal. The image recognition method based on cloud analysis and processing needs to consume a lot of communication resources, cloud computing power, storage capacity and other resources, and its real-time performance is poor and the cost is high. Therefore, the image recognition method gradually shifts from the clo...

AI Technical Summary

Problems solved by technology

[0003] At present, smart terminals with edge-end image recognition capabilities are relatively lagging in the introduction of evaluation specifications, and the evaluation technology and device compatibility are poor, and the evaluation pressure is huge.

Taking the intelligent visual monitoring terminal as an example, its image recognition ability evaluation methods mainly include the following two types: (1) model extraction type ability evaluation, this method decouples the model to be tested from the terminal, and can verify the effectiveness of the algorithm itself, but It is impossible to detect the effect of the model in the actual application of the terminal; (2) manual import capability evaluation, which manually imports test cases through SD card or mobile storage tools, and cannot directly call the sample library, resulting in low test efficiency. The method also cannot evaluate the power consumption of the terminal

Images