Server component intelligent detection method based on deep learning

Abstract

The invention discloses a server component intelligent detection method based on deep learning. The method comprises the steps that S10, a sample image is collected and marked; s20, feature enhancement is performed on the acquired sample image; S30, a target detection model based on a neural network is constructed, wherein the target detection model comprises a feature extraction network and a multistage feature fusion network, image feature information is extracted through the feature extraction network, and a prediction matrix is constructed through the multistage feature fusion network; S40, a Focal Loss loss function is constructed; S50, a target detection model is trained by using the enhanced image data, wherein after each iteration, the weight of the next iteration process and the updating trend of the offset are determined according to a Focal Loss loss function; and S60, the trained model is used for detecting the server components. According to the server component intelligent detection method based on deep learning, the positions and the number of the server components can be accurately recognized, and the server component assembling efficiency and accuracy are improved.

Description

technical field [0001] The invention relates to the technical field of industrial visual inspection, in particular to an intelligent detection method for server components based on deep learning. Background technique [0002] Manual detection has problems such as labor-intensive, human visual fatigue can easily cause detection errors, is affected by subjective factors, and has low accuracy. The traditional detection processing method of industrial vision accessories is to extract the connected domain, match it with the preset area, and compare the results. Usually, in the process of extracting connected domains, images are easily affected by light sources and noise, and correspondingly, the matching results will be greatly reduced. In addition, this processing method is relatively rigid. In a more complex industrial environment, problems such as the deviation of the angle of the accessories and the change of the working distance will affect the detection results. Therefore...

AI Technical Summary

Problems solved by technology

Usually, in the process of extracting connected domains, images are easily affected by light sources and noise, and accordingly, the matching results will be greatly reduced

In addition, this processing method is relatively rigid. In a more complex industrial environment, problems such as the deviation of the angle of the accessories and the change of the working distance will affect the detection results.

Therefore, the robustness of such a detection method is poor, and the error rate of the detection result is high.

Images