A fully convolutional neural network-based method for water level detection and validity identification of water gauge
A convolutional neural network and recognition method technology, applied in the field of water level detection and validity recognition of water gauges, can solve the problems of inability to recognize validity, too much background interference noise, low image resolution, etc., so as to facilitate manual verification, The effect of high detection accuracy and high degree of automation
Active Publication Date: 2022-07-29
HOHAI UNIV
View PDF2 Cites 2 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, in practical applications in the field, due to the long distance between the water gauge and the shooting equipment, the image resolution is low and the background noise is more. The glare on the water surface and the reflection of the water gauge are strong, and the gray values of the water gauge and the water surface are close to each other in rainy weather, which makes it difficult for the conventional image processing technology based on grayscale and edge information to extract the water level line, or the detection error is relatively large. The validity of the measurement results such as the water gauge is blocked by aquatic plants cannot be identified
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 Construction
[0054] like figure 1 As shown in the figure, a method for water level detection and validity identification of a water gauge based on a fully convolutional neural network includes the following steps:
[0055] S1. Obtain the water ruler image and perform manual labeling, and distinguish the water ruler, aquatic plants and water bodies (background) with different categories to obtain a label map;
[0056] S2. Design a fully convolutional neural network structure and conduct network training;
[0057] S3, using the fully convolutional neural network obtained by training to perform semantic segmentation on the image to be tested to achieve pixel-level semantic labeling;
[0058] S4. Detecting the water level of the water gauge in the semantic segmentation image and identifying the validity.
[0059] The acquisition of the water gauge image described in step S1 includes the following steps:
[0060] S1.1: Select 24-bit actual water gauge monitoring image data under different w...
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 invention discloses a method for detecting and validating the water level of a water gauge based on a full convolutional neural network, comprising the following steps: S1. Obtaining a water gauge image and manually marking it, and classifying the water gauge, water plants and water bodies into different categories Distinguish, get the label map; S2, design a fully convolutional neural network structure, and conduct network training; S3, use the fully trained fully convolutional neural network to perform semantic segmentation on the image to be tested to achieve pixel-level semantic labeling; S4, in semantic segmentation Detecting the water level line of the water gauge in the image and the validity of the recognition. The invention has high detection precision, strong robustness and simple operation.
Description
technical field [0001] The invention relates to the technical field of computer vision, in particular to a water gauge water level detection and validity identification method based on a fully convolutional neural network. Background technique [0002] Water level is one of the basic hydrological elements of rivers, lakes and reservoirs, and an important indicator to reflect changes in water bodies and water flow. Water level data is the basic basis for the construction and management of flood control and drought relief, irrigation, shipping and water conservancy facilities. Water resource management planning and sustainable development are of great significance. The water gauge records the height of the water level by reading, which is the most intuitive and simple measurement tool; however, the traditional water gauge measurement requires manual regular observation, which has a low degree of automation and high labor intensity. The existing automatic water level gauges ma...
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 Patents(China)
IPC IPC(8): G06V10/26G06V10/774G06V10/764G06V10/80G06V10/82G06K9/62G06N3/04G06N3/08G06T7/00G06T7/136
CPCG06T7/0002G06T7/136G06N3/084G06T2207/20081G06T2207/20084G06T2207/30204G06T2207/20221G06V10/267G06N3/047G06N3/045G06F18/2415G06F18/241G06F18/253G06F18/214
Inventor 张振周扬王慧斌张丽丽汪崎宇李嘉辉沈淏旸
Owner HOHAI UNIV