Intelligent collection method for machine room equipment display panel data

Abstract

The invention discloses an intelligent collection method of machine room equipment display panel data, which comprises the following steps: S1, a robot acquires images of a plurality of display panelsas a training data set; s2, inputting the training data set into an improved master-rcnn algorithm, and training to obtain a text detection model; s3, the robot collects an image of the display panelin real time and inputs the image into the text detection model obtained through training in the step S2, all texts are automatically marked, so that detection boxes are obtained, and position coordinates and size information of all the detection boxes in the image in the display panel are output; s4, extracting a Roi image in the detection frame, performing image preprocessing on the Roi image,and reserving the extracted digital skeleton image as a training sample set; s5, training an svm classifier according to the training sample set, and performing classification recognition on a singlenumber through the svm classifier; and S6, splicing the numbers into a character string, and outputting the character string to a client for display. Data is automatically collected, the labor cost isreduced, and the operation and maintenance efficiency of the data center is improved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to an intelligent collection method for display panel data of equipment in a computer room. Background technique [0002] With the rapid development of artificial intelligence in today's era, traditional industries are gradually seeking to transform from labor-intensive to automated. There are a large number of IT equipment installed in the computer room of the data center, including servers, network equipment, etc. These equipment are equipped with various display panels, and it is necessary to manually record the panel readings around the clock to check the current operating status of the equipment. Cost, and the accuracy of data collection cannot be guaranteed, and it is prone to reading or transcription errors. Manually transcribing data around the clock is extremely boring and inefficient. Modern measurement requires automation as much as possible to save manpower an...

AI Technical Summary

Problems solved by technology

[0004] (1) Due to the wide variety of display panels in the computer room, the relative positions of the numbers in the panels are also different. To determine the position of the numbers by manually marking the digital area, the operation is complicated and cannot realize real intelligence and automation, and the accuracy of the marking Susceptible to subjective factors, thus affecting the final recognition result

[0005] (2) Since traditional methods such as image preprocessing and projection methods used in the existing technology rely on the effect of image binarization, and preprocessing operations such as binarization are poorly adaptable to illumination, the impact of illumination changes on the final recognition result bigger

[0006] (3) Consider using other text recognition methods such as the mainstream OCR recognition library, but the effect is often better for printed panels, and the seven-segment digital tube number itself is composed of seven small segment digital tubes, and the digital tubes are not glued. Existing OCR library recognition is easy to recognize small segments of digital tubes individually, but cannot recognize the overall digital meaning, and the recognition accuracy is low

[0007] Therefore, the intelligence degree of the existing method is low, and it still inevitably needs manual assisted identification, and it is difficult to meet the automation requirements. Aiming at the digital display panels of the seven-segment digital tube type that exist in a large number of data center computer rooms, the present invention proposes a computer room equipment Intelligent Collection Method of Display Panel Data

Images