Presentation is loading. Please wait.

Presentation is loading. Please wait.

Hydrology for all: Image-based water level monitoring.

Similar presentations


Presentation on theme: "Hydrology for all: Image-based water level monitoring."— Presentation transcript:

1 Hydrology for all: Image-based water level monitoring

2 How does it work? Finds water line in an image Calculates the water stage Stores both the value and the image Provides live services in the form of stage, images, videos, alerts.

3 What does it take? A good Webcam A good camera stand A good flat background An ftp server A machine vision software to calibrate the images Remote clients (smartphones, computers, etc.) to visualize, download data

4 A good Webcam Objective: install virtually anywhere in the world Live images require 3G coverage Picture of unsecured prototype: new packaging in testing

5 A good 3G webcam

6 A good stable stand for stable images

7 Picture of older prototype. New and improved prototype in testing

8

9 A good flat background

10

11

12

13 A machine vision software to calibrate the images

14 Remote clients to visualize, download data

15 Preliminary results Results borrowed from – Gilmore, T.G.,F. Birgand and K. Chapman. Source and magnitude of error in an inexpensive image- based water level measurement system. Submitted November, 10, 2011 to Journal of Hydrology

16 Can we detect water level?

17 Can we measure water level? Detection and linkage to real world measurement

18 The System: Edge Detection

19

20

21

22 The System: Calibration

23 x = 0 cm, y = 75 cm x = 0 cm, y = 45 cm

24 Uncertainty: Three Experiments 1.Image Resolution 2.Lighting effects* 3.Perspective 4.Lens distortion 5.Water meniscus Benchmark II Benchmark I Water Level

25 Uncertainty: Camera effects Lighting* Meniscus Effects

26 4m, 12 mm5 m6 m7 m cm per pixel Mean Bias, SD, RMSE (cm) Standard error for several image resolutions in the lab

27 Water Level: Posture Angle 19 deg 10 deg For camera at 6 meters

28 Water Level: 6m, 16mm lens posture angle from horizontal (degrees) 1910

29 Preliminary recommendations and performance 1.Lens distortion must be minimized 2.Posture angle may interact with meniscus 3.With reasonable precautions, accuracy of +/- 3 mm (0.01 ft) is achievable in the lab

30 Products Data Pictures Videos Alert Remote actions in the field (work in progress)

31 Miscellaneous pictures

32

33

34 Picture of Prototype of camera mount – a secured version is in testing

35

36

37

38

39

40 Please visit


Download ppt "Hydrology for all: Image-based water level monitoring."

Similar presentations


Ads by Google