Transformer substation safety fence crossing behavior identification method, system and equipment
A security fence and recognition method technology, applied in the field of image recognition, can solve the problem of low accuracy of behavior recognition and achieve the effect of improving accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 2
[0064] Take the video of workers climbing over the fence and the video of workers walking normally. The videos contain different weather conditions and different lighting; the video of workers climbing over the fence is used as a positive sample, and the video of workers walking normally is taken as a negative sample. Based on positive samples and negative samples Sample construction video data set D1; construct video data set D1 by obtaining positive samples and negative samples, so as to train the deep learning network in the future;
[0065] It should be further explained that the specific process of constructing video dataset D1 based on positive samples and negative samples is as follows:
[0066] Mark the positive sample and the negative sample respectively. Since the positive sample is a video of a worker climbing over the fence, the positive sample is marked as crossed. Since the negative sample is a video of the worker walking normally, the negative sample is marked as...
Embodiment 3
[0097] Such as Figure 4 As shown, a substation safety fence climbing behavior recognition system includes a video data set module 201, a skeleton point image sequence collection module 202, a motion stream video sample set module 203, an optical flow video sample set module 204, and a deep learning network training module 205 And a real-time behavior recognition module 206;
[0098] The video data set module 201 is used to obtain the video of the worker climbing over the fence as a positive sample, obtain the video of the worker walking normally as a negative sample, and construct the video data set D1 based on the positive sample and the negative sample;
[0099] The skeleton point image sequence collection module 202 is used to extract the human skeleton from the video data set D1 using a pose estimation algorithm to obtain the skeleton point image sequence collection K1;
[0100] The motion stream video sample set module 203 is used to perform tensor voting on the human s...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


