Question

I have to develop an Intensity Meter, which will basically display the average intensity level of a uniformly exposed image. As the focus of the camera is changed the pointer starts moving to the correct value in the following display:

Along with that a text field will be displaying the exact intensity value in digit as well, as can be seen in the figure.

Till now I have been able to capture the image and store its intensity values in a text file. Now I have to develop this animated image. I have never done any animations using OpenCV, so I am looking for some idea how to go about this kind of animation using OpenCV.

Any pointers here?

My complete application is based on Windows Forms (C++) and I am using OpenCV for other
Image processing tasks (not listed here).

Note: The meter will basically show average intensity level of a uniformly exposed area.

Update

I have got some solution here.

So with some research I now know the algorithm:

Theta= A * I

Where,

Theta= angle at which the pointer is rotated

A = Proportionality factor

I= Intensity Level

The angle of rotation will be directly proportional to the intensity level.

So now here is how I plan to go:

1- Create a new Window using cvNamedWindow

2- Display the static part of the image , i.e the dial in the figure, which never changes

3- Display a vertical Pointer pointing at the middle

4- Get the value of Intensity Level I, and apply the formula above to calculate Theta.

5- Based on the angle calculated above, rotate the Pointer by using OpenCv Function.

Can some verify my understanding? Especially please tell me what is the fastest function for rotating the pointer? Please let me know if you can improve it further.

Answer 1

I’ll suggest two options.

1. Use a library: something like WxIndustrialControls which I linked in the comments.
2. If you really want to draw things in opencv, please post the code that you have so far so that answers can focus on the part that has you stuck.

I haven’t done drawing in opencv, but I think I found the parts you would need:

• simple image load (e.g. imread or cvLoadImage)
• drawing functions (e.g. fillPoly to draw your needle
• seems like you can update images quickly using simple cvShowImage as this person did with a camera feed
• an equation that converts your input intensity value to a set of points for fillPoly

Please update the question with more details about which parts have you stuck. Sorry if I linked to multiple versions of opencv documentation. Shouldn’t matter for these fundamental functions I guess.

update for the equation:

I haven’t tested this at all, but it seemed like something fun to come up with:

Assumptions:

• left of center ==> 180 deg
• right of center ==> 0 deg
• image origin is top left
• pixel order is [column, row] (i.e. [x, y])

Configuration Variables:

• min_v ==> the value to go on the left edge of the dial
• max_v ==> the value to go on the right edge of the dial
• arc_size ==> the number of degrees at the top of a circle that you want to use for your intensity range
• radius ==> the distance from the top of your dial to the center of a virtual circle (larger radius ==> flatter dial)
• top_padding ==> the distance from the top of the static image to the needle when pointing straight up
• needle_l ==> the length of the needle / polygon you want drawn

Input Variables:

• needle_v ==> value for the needle to represent

Output Variables:

• needle_end ==> the outer point of the needle in terms of your image coordinates
• needle_start ==> the inner point of the needle in terms of your image coordinates

Calculation:

for readability:

left_a =  90 + (arc_size / 2)             (angle for min_v)
right_a = 90 - (arc_size / 2)             (angle for max_v)
value_range = (max_v - min_v)

get the angle for the needle

value_proportion = (needle_v - min_v) / value_range
needle_a = left_a - (value_proportion * arc_size)

get the virtual center of the circle (e.g. the center for the one in your example image would be outside the image)

dial_origin = [int(image_width / 2) , top_padding + radius]

get the needle start and end points

needle_end[x] = dial_origin[x] - radius * cos(needle_a)
needle_end[y] = dial_origin[y] - radius * sin(needle_a)
needle_start[x] = dial_origin[x] - (radius - needle_l) * cos(needle_a)
needle_start[y] = dial_origin[x] - (radius - needle_l) * sin(needle_a)