Good question.  We haven't posted any practical code examples with the new C interpreter, so this is an ideal opportunity.  We might try to use the lasers and blob detection as a starting point.  I will work on this tomorrow, though others are certainly welcome to jump in.

There are two parts to the problem -  
 
1.  an obstacle detection algorithm that uses sensors and processing
2.  a motion control algorithm that applies the results of obstacle detection
 
The main issue with reflected laser light is that it can be confused with other light sources.  If you are confident that the reflected laser light will produce the brightest pixels with the camera, then you can search directly for laser pixels.  Otherwise, the most certain way to resolve this is to capture a frame with the lasers off and compare it with a frame with the lasers on.  As long as the robot or background scene isn't moving, the only image left after subtraction should be the reflected laser light.
 
A good way to test this is to use the SRV1Test application to sample different color ranges and visually observe the results.  From the SRV1Test command window, send
 
'l'  (lower case L) to turn on lasers
'g1' to view color segmentation results
'vc0250255100150100200' to set color bin #0 to Y [250-255], U [100-150], V[100-200]
 
If you only see laser light reflection and no other noise, then you've created a good range of colors.  You can confirm by running the blob command
'vb0'  
and you'll see the coordinates of the reflected light color blobs.
 
You can perform the same functions in the  C interpreter.
 
laser(3);  // turns on both lasers
vcolor(0, 250, 255, 100, 150, 100, 200);  // sets up color bin #0 as above
vcap(); // grabs a video frame
vblob(0,0); // searches for the largest blob matching the color bin #0
 
vblob() actually returns a value which is the number of blobs it finds.  Generally, we're interested in the largest blob, but with two lasers, we should get two blobs, so you might want to look at the details of both blobs.  The results of the blob search are stored in global variables  
    blobcnt (number of pixels in the blob)
    blobx1 (left-most x coordinate)
    blobx2 (right-most x coordinate)
    bloby1 (top y coordinate - 0 is the top row)
    bloby2 (bottom y coordinate)
 
With that information, you have your blob location, so then you have to decide what action to take.  For starters, we should put together a small program that just turns on lasers, does a blob search, and dumps the result.  
 

Jim -
 
I would be interested in your feedback on how we might best use the edge detection function in the picoC environment.  Along those lines, I just posted a note -  
    http://www.surveyor.com/cgi-bin/yabb2/YaBB.pl?num=1241199304
about possibly adding a vfind() function to search a region.  Perhaps we should also add vedge(thresh) to perform the edge filtering and then use vfind() to count the pixels that have been flagged as edges ?
 
The issue with line lasers is that the reflected light is not very bright unless you get a laser with fairly high output.  I agree that it is best to work with image features, though adding a visible light source to the robot can help provide a more stable vision processing environment.   That's actually something we have in the works.

Good suggestions in your other thread about employing edge detection -
       http://www.surveyor.com/cgi-bin/yabb2/YaBB.pl?num=1241221826
 
I will give more thought to the functions in picoC that dbibb needs to accomplish the stated task.

Yes - that is the purpose of the signal() function - it detects if any characters have been received from the host and can be used to break out of a loop.

Certainly.  You can put multiple LED's in series, as the voltage drop should be 0.7V per LED.  You can put them in parallel as well, but watch your overall current draw.  There are some very bright LED drivers available as well - Dimension Engineering sells one, and we're aware of another becoming available soon.

Those sample images really show the value of having a light on the robot.  
 
My friend came up with a fascinating idea.  He proposed using one of the new micro-projectors instead of a lamp.
 
They might not be bright enough yet, but the idea having a controllable light source, that could optionally project a grid, white light, or even a map of some kind, is pretty cool.
 
Something to think about.
 
Jim