|
Colorado and Oz | Great combo - a flying camera and an AgAerialCamera aka Aerial AgCam (AAC). I will be a follower of your progressions. Tell us more!
You likely know this but the best resolution of information question continues in this next phase of precision farming systems. The imagery for best matching by filter/channel to precise field position, by management grid, a swath corridor, row, or plant is open. The precise aggie was excited and confounded to mange at acre to hectare scales and now insight over a crop-space smaller by a factor of 1,000. Spot-spray bugs and weeks before the rascals get away?
SO what may be the best remote sensing resolution for ag:
To test for seedling stand having soil surface camera resolution better than 4 inches or 100mm may not be adequate as discussed? To get one inch resolution you would either require a larger sensor in pixels width and lenght. Alternatively you could increase length of the lens... you get a smaller more narrow field of view of the ground by the lens zoom effect - same number of pixels on a smaller field of view due to the longer lens.
To test for irrigation efficiency, 4-inch (100mm) infra-red resolution might require a $20k camera and a $10k platform to lift it... when a review of your irrigation set with a lower res IR camera, say at three feet per pixel (smaller and/or higher), may cost a fraction of the high-res and yield exactly the same management result.
At these scales camera innovation will be savage. Old-hands will help others rediscover the hard-knocks of nadir (straight down) versus oblique (looking at an angle) perplexed by image distortion, cloud modeling and stitching and yet fully shared best practices discovered for our various terrains, Beyond what color do you fly will be conversations on lenses, rolling vs global shutter, and all sorts of synchronizations and grid correlation study . Sounds simple but getting perfect nadir is not as simple as you will assume. Good software and plenty of image to image overlap can sort most of the non-nadir-ness out? How free can it be?
To be able to calculate the resolution of one pixel = what area at the ground you will need to know: the sensor size typically in mmh x mml, the pixel pitch or pixel size, the lens in terms of its focal lenght also in mm, Other factors to see how far your 20 minute flight will get would include average speed-over-ground, the AGL (above ground level) altitude of the camera, and desired overlap percentage. Just as the soil sampling gridding scale varies depending on the control-ability of the input versus measured variation of its prescription, so will the precision AgCamera: that more spatialized information tends to confirm un-controllable fine variation at greater expense, expect the same for AAC. If you are making "video" then you trade many frames at lower resolution.. even if you claim 1080 frame size. A 1080 video frame is at best low res at about 2Mb per image depending on the video compression. If a picture camera is selected many more pixels per frame are achievable but the frequency of the next frame may be a second or two separated.
If you have not discovered them there a several mini-cams that split a single aperture into four or more filtered channels. If you split a 1080 video frame you will get four fully synchronized at aperture, by frame, and lens. If you use a 4K camera you will get four 1080 frames. Lots of trade-offs to be discovered...
What was a satellite or manned flying camera at $30,000 per channel or far more "last year", may be less than a $1000 likely to drop to half that next year? Looking forward to seeing more for less...
MidNight Mapper
aka neil
Edited by MidNight Mapper 6/12/2016 10:54
 (Parrot_Multispectral.jpg)
Attachments
----------------
 Parrot_Multispectral.jpg (78KB - 437 downloads)
| |
|
Best images for ag?
