Training method and detection method of electric wire based on depth separable convolution neural network
A technology of convolutional neural network and training method, which is applied in the field of image processing and pattern recognition, can solve problems such as the inability to guarantee that the network has learned the characteristics of wires, and achieve the effects of strong interpretability, stable effect, and flight safety
Active Publication Date: 2019-03-29
SHANGHAI JIAO TONG UNIV
View PDF2 Cites 11 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0007] 1) If you directly use the network to output whether there are wires on the picture, you need to limit the size of the network input, which will not only scale the picture and lose information, but also cannot guarantee whether the network has actually learned the characteristics of the wire
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0047] This embodiment provides a wire training method and detection method based on a small-scale separate convolutional neural network. The general idea includes 4 steps in the training part and 5 steps in the detection part:
[0048] Training part:
[0049] S1, building a training network;
[0050] S2, prepare the grayscale single-channel image of the wire, and the corresponding binary mask truth map;
[0051] S3, grouping 13×13 slices centered on each pixel;
[0052] S4, input the slice into network training to obtain network weight parameters.
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The training method and detection method of electric wire based on depth separable convolution neural network comprises that steps of: constructing a neural network using the deeply separable convolution; trainingconvolution kernels with uniformly distributed mini-slices, wherein the trained convolution kernel is used for feature extraction of infrared gray-scale image; binarizing the image according to the threshold value, removing the small area region, and connecting the linear region by the probabilistic hough transform. The invention trains a convolution kernel capable of extracting infrared wire features by a machine learning method, and can effectively extract wire features from infrared gray-scale images. Combined with morphological processing and linear Hough transform, wire detection can be carried out in real time.
Description
technical field [0001] The invention relates to the technical field of image processing and pattern recognition, in particular to a training method and a detection method for electric wires based on a deep separable convolutional neural network. Background technique [0002] Safety issues in helicopter flight have always been a hot topic in the industry. Existing helicopter obstacle avoidance mainly relies on manual visual observation, limited by the resolution distance and resolution ability of human eyes, small obstacles such as high-voltage wires are not easy to be detected. In addition, the pilot needs to concentrate very much to find obstacles such as high-voltage wires ahead, which greatly increases the work intensity and mental stress of the pilots. The helicopter automatic obstacle avoidance system based on optical, radar and infrared technologies can improve the flight safety obstacle avoidance ability of the aircraft under low visibility conditions. It is an advan...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/46G06N3/04G06N3/08
CPCG06N3/08G06V20/20G06V10/462G06N3/045
Inventor 李元祥刘嘉玮龚政庹红娅周拥军
Owner SHANGHAI JIAO TONG UNIV
PatSnap Eureka turns technology decisions into work you can execute. Powered by our Innovation Knowledge Graph, it runs expert workflows across engineering, life sciences, materials and intellectual property. Get your review-ready output in minutes.
