I will take my proposed aerial photo survey one step further.
Rather than measure the body of water directly, one could mix a calibrated amount of a non-toxic environmentally neutral dye in with the chemical treatment. Then fly a drone over the rice paddy to determine that the field has the the right color which would indicate the appropriate measure of application.
Hmm.. a quadcopter would blow a lot of air down on the water surface unless it's high above, wouldn't it? And the farther away, the more difficult it would be to measure the water depth I imagine - and I can't help it would be difficult in the first place, due to the air/water interface.
Boats may also run into issues, the water depth is small to start with and sensitive to displacement effects.
I'm sure that the simplest possible approach would work the best.
In the ocean, the propagation of water waves and their shape can be dependent on depth of the water. Perhaps this phenomenon can be extended to very shallow water. If you have a movable device in the water that produces waves with certain characteristics and then some sort of camera or optical system to measure how the water waves move and dissipate, maybe you can deduce the depth of the water.
The biggest problem, however, is all the localized mins and maximums in the mud, and these add up, as has been mentioned earlier. You can assume they average out but I doubt that's a good assumption - they will probably err on the side of going deeper than the apparent average.
Of course, once you get to a certain point with all this technology, it would probably be cheaper to build an army of robots that march through the mud, hunt down and kill off the snails one by one.
I do admit that aerial survey is not without technical problems, but the amount of time and effort that would have to be spent on survey is a lot less. So it would seem that through developing a good method of measurement and calibration method, it would be easier to deploy. Reflected light from the water surface does take on the color of the sky, but there is always the option you use polarized light filters.
Optical calibration can be done in the field with actual samples of water in the particular field inclusive of sediment on the day of the test and chemical application. This doesn't have to have perfectly clear water or a windless day. The morning usually has less wind, so there may be a particulary time of day that is optimal for using this method.
I've been asked to come up with a method to figure out how much water there is in a rice paddy just before planting. This has to do with voracious snails that eat the shoots, and biocontrol using a soapy substance extracted from quinoa hulls. The pesticide is expensive, and the dose per volume needs to be well controlled. Each paddy covers about 4 hectares (200m x 200m), and the water is about 4 to 7 cm deep. The way they do it now is to have a couple of people walk around in the paddy with a meter stick, and while one takes a reading, the other writes it down in a notebook for later transcription to a computer for the calculation. They take several samples at the corners and at the center. The accuracy is not good.
I do like all the wheeled robot suggestions. A few bike wheels and car window motors would work great. A floating "tail" would then easily probe the surface height. Also, you could do some dead-reckoning with the wheels to augment the accuracy of the GPS. And the final product would be a convenient platform for other samplers and sensors. (i.e. soil chemistry, etc)
Of course the high-tech method would be to use a large area 3D scanner to scan the height of the field just before it was flooded.
Spiking the pesticide with a crop-safe dye is also interesting. You could then optically monitor the concentration of the dye for a closed loop pesticide dosing. i.e. hit the field with half the expected dose, check the concentration some hours later, back calculate the water volume in the field, and adjust the final dose to suit. Thinking further, just dosing with a crop-safe dye and measuring the final dye concentration should allow you to measure the water volume in the field. (assuming negligible flow out of the field and uniform mixing...)
Just thinking out loud ... I wonder if something along the lines of a mini Tsunami where assuming that the water is deeper in the middle than on the peripheral, a displacement 'wave' could be produced in the middle and measured along the edges.... i.e timing, magnitude etc. Calibrating that to equate to volume might be tricky, but I think it would be consistent.
A tsunami, making waves! That could be interesting, the interference effects, the propagation velocity.
However, the first step has to be something that can be considered a "gold standard" against which others can be measured. I'm liking the wheel robot too, with the sensor on a spoke or linkage to a float or submerged tail. Wheels for hand trucks are 25cm (10") diameter. A tricycle for stability, or just float the trailing end?
The idea of a dye is intriguing, for back-calibration. It brings up the question of how one goes about disseminating the stuff and mixing it into such a large flat sheet of water. Spray? Tractor? I'll have to find out. Water is lost via percolation into the ground and transevaporation into the air, but the dye marker and the pesticide would have the same concentration dynamics. During the application, all intentional inflow and outflow will cease.
If the water is clear enough for Phil's laser idea to work, why couldn't you use his idea using a 2D sensor instead of a line sensor?
Have a laser up on a pole projecting a line running straight across the field in front of a camera. The camera could then be used to measure the distance between the line on the top of the water and the line at the bottom of the water.
The line will of course be effected by the refraction of the water but I'd think this could be corrected for (I''m pretty sure with the geometry shown, the water will cause the two lines to appear closer together).
I liked this idea better previous to making the above illustration. I think the distance between the two lines will be hard to measure for all but the shortest distances. The camera's position and orientation would have to be very accurately known for this sort of setup to have a chance of working.
Edit: I've asked Graeme Finch if it's okay to use his photo this way. I haven't heard back from him yet but if he objects, I'll delete it.
I'm liking the wheel robot too, with the sensor on a spoke or linkage to a float or submerged tail. Wheels for hand trucks are 25cm (10") diameter. A tricycle for stability, or just float the trailing end?
If you float the trailing end, you might be able to get depth from deck angle with an accelerometer, eliminating pivoting arms or float tubes. Torque effects might make a tricycle arrangement more convenient, however. WRT wheels, you might have a look at these: http://www.gokartsupply.com/tires.htm. If you scroll down a bit, they have what looks like mini-tractor tires which ought to do well in mud. Go kart or scooter supply companies should have lots of tires to choose from, and wheels made for accepting drive sprockets.
I think you're going to need a very wide, very smooth wheel: wide, so it doesn't sink in the mud; smooth, so you have a known, fixed ouside diameter. If you don't get enough traction from such an arrangement, you may need to trail both sinking and floating probes and measure the angle between them.
This is a situation where I don't have enough personal experience in rice fields to say what's what. I looked up "paddy tractors" based on comments above, and found big lug tires, deep tread, and here is a fragment of a photo from <http://www.fotopedia.com/items/dbpo90it15iml-vhmr4-ltfh8>. Big lugs, spokes, no tire in between! Quite the opposite of smooth. This suggests to me a very slippery substrate on top of a firmish base.
The Go Kart place has some big-lug tires, pneumatic. I'd been thinking more along the line of a solid rubber wheel with plastic hub as you would find on a cheap hand-truck. But those are flat, not much tread.
The mud layer holds some fraction of the water and for that reason may need to be counted, so the deep tread may be a feature. Can't over-think this!
Almost any wheel could be modified to work on muddy bottoms as long as they do not float. Put nails or screws through them for traction. Concrete nails work really well. Could even cast concrete wheels with the spikes included. We put screws (with nuts) through the tires for racing go-karts on ice.
The rice fields aren't that muddy really, there shouldn't be any need for wide wheels as long as you're not using heavy machinery. People are walking in the ricefields when planting, after all. And herons looking like they must weigh a couple of kilos at least walk around with no indication of getting their thin, skinny feet sucked down.
I can't help but think the geological diversity of rice paddies is huge as rice is grown on nearly every continent, excepting Antarctica.
Speculation about not being muddy is dependent on local soils and local water sources. Soil is generally classified as a mixture of three non-organic elements -- clay, silt, and sand. And then there are the organic features suspended in the water.
I am beginning to think a narrow probing wheel that runs along the bottom, with a float that records the actual water surface would be best.
Why so? The weight of the robot, inclusive of batteries would get the bot resting on the bottom. It is much easier to monitor the water surface than to have a probe reach the actual bottom with any certainty. Mass would be aligned with a low center of gravity and the weight of batteries would be actually working for you.
I'd still rather do it all with a fly-by survey and not have to run machines over the surface. Machines can fail in many ways and have a lot of down time. And there always is a tendency to revert to gasoline driven machines in such a context.
This begs a site visit to get a better feel for it. Not too far from Parallax HQ is California's rice growing region on the Sacramento River delta.
Duane, that is quite a picture in #39, with the laser tower on the side. I too hesitate to use someone's online photo, what constitutes fair use? I figure a link to the source is the least we can do.
I do hesitate about any optical method. Roughness or undulations or water turbidity could kill it.
Now looking at the path that the rice robot, aka paddy wagon, would have to follow. The perimeter alone is around 800m, and a jaunt to the center brings it up to near 1km. What would be a reasonable speed? Let's suppose 0.25m/s. That means somewhat over an hour to run the course. That is acceptable. But is that a reasonable speed? I attached an animated gif of a possible path that would cover even more points and bring the bot back to its starting point.
What are the chances of any survey device getting stuck in the middle of the rice paddy? And if the weight is more than one man can easily pack for 100 meters or so, how are repairs and recharges going to be handled?
This is farm machiery and will need to be fully capable of field repairs and troubleshooting. And don't expect it to not be muddy and wet... expect the unexpected.
I suspect a floating support vehicle would be nearly a necessity to carry tools and new batteries.. maybe to pull the machine out if necessary.
I think you'll get a more accurate and easier-to-compute volume estimate with a trajectory that visits regularly-spaced, orthogonally-arrayed sample points, such as a back-and-forth or square-spiral traverse. Can we assume that the bot will be GPS-guided?
Edit: On second thought, I reread the requirements. 4cm - 7cm is very little room to float in, so I don't think this device will work.
I favor simplicity. I propose a stick that has a weight at one end, a float at the other, and the electronics inside. The water level is gauged by the angle of the stick and measured with a three axis accelerometer. The position of the stick is determined by a GPS in the float.
In order to cover a whole field the stick has a string attached that can pull the unit to the next location. So, the farmer can easily drag the stick across the field. I've also thought it would be useful to shoot the unit from a bow or ballista.
The advantage here is simplicity: no moving parts. You could also make a bunch of these for very little (<$100 parts) and parallelize your field exploration, or leave one behind to monitor water level over time.
Here's a graph of acceleration as the measurement device goes from vertical to horizontal:
It's more than sensitive enough to measure small changes.
Phil, the device would be GPS equipped. I agree, a rectangular back and forth or spiral grid would be best. I see it taking continuous samples as it moves around, one or a few per rotation of the wheel. It may have to move, stop, move, stop, to let the wake die down.
That's an interesting idea, Cody. I could see a couple of guys on opposite sides of the pond with a loop of fish line in between. They pull the ducky across or have a motor do it at constant rate, and then move it a few meters over. Like striping a football field. What device did you use to make the graph? The fact that it is sealed and does not need direct water contact is good point.
Loopy, I too have stories about things that go bump in the night or stick in the mud for outdoor and farm installations. Enough to be cautious and skeptical. Just brainstorming. It's too early to think about repairs and maintenance costs on the vapor!
Tor, I value your observations about the rice paddies. In Japan now? You must have a very nice walk to work. One of my favorite scenes in cinema is the rice planting to song at the end of Seven Samarai. The surviving Ronin observe that the farmers are the victors.
Of course, with laser guided earth moving equipment, it might just be easier to go in and level the paddies to the point where you don't have to measure the uneveness for a few years. By then, the snails are all gone and the fields can revert to their prior chaos.
There must be someone in the Department of Argriculture that has looked at this issue before.
Rice growing in California is very interesting. They have eliminated the need for application of chemicals to their rice paddies by not growing rice for the table. I was told that the entire state's crop is sold to Budweiser for making beer as it doesn't have to be cometically white and beautiful.
You might argue that they need to apply fertilizer, but the grower I was talking to would take all the rice stalks to Petaluma and use them for bedding in his chicken farms. And then the bedding with chicken poop was later plowed back into the fields. By producing rice and chickens and selling for beer, the rice growers avoid a big chunk of costs related to farm chemicals.
Taiwan grows rice. In fact it grew about 40% of Japan's rice during WWII. And it has the snail problem everywhere. Eradication has been a huge failure. I am not sure what they do to mitigate. Thailand is a world leading exporter and rice producer. It would be worthwhile to find out how they manage the snail problem.
Japan's rice farming is heavily protected and subsidized by the Japanese government. I am not sure that anything they do reflects economically compeitive reality.
Japan's rice farming is heavily protected and subsidized by the Japanese government. I am not sure that anything they do reflects economically compeitive reality.
That's an interesting observation. Some countries have food commodities that they identify with and that are nearly sacred. The same applies to Mexico's corn tortillas, which are also federally subsidized.
_________________
Tracy,
In a good 3D CAD program like Rhino, you can import a "point cloud," convert it to a surface or mesh, then have the program compute the resulting volume -- all without having to write a line of code.
I like the Peano curve. It is scalable.. doesn't have to be so detailed.
I wonder about the affects of wind on 3" of water. The windward side of the a paddy might only be 1" deep while the other side is 5" deep. The hazard is that you start measuring without any wind, and then halfway through the wind picks up and distorts the results.
It just might be the best alternative to rid the area of snails would be a few hundred hungry ducks marched through each paddy at the right time. No chemicals at all. And more revenue from all that duck meat and those duck eggs.
Tor, I value your observations about the rice paddies. In Japan now? You must have a very nice walk to work. One of my favorite scenes in cinema is the rice planting to song at the end of Seven Samarai. The surviving Ronin observe that the farmers are the victors.
I came back to Norway recently, will return to Japan soon. Watching the changing rice paddies is indeed nice. One of the most photographed features of Japan and some other rice-growing regions must be the terraced rice fields in the spring, when they're full of water. Well, those types at least won't be easy to manage with any kind of machinery..
Love visiting Japan... especial Central Honshu, the samurai castles, the shores of the Sea of Japan, and the Japanese Alps. But planted rice paddies are almost hypnotic anywhere.
It seems like each section of a rice paddy must be level so it evenly fills with water. Just measure the depth at one point and multiply times the area.
EDIT: If the rice paddies aren't level then take a series of aerial pictures with various amounts of dry land exposed, and measure the water depth at the deepest point. Determine the amount of covered area for each depth measurement and you can then compute the volume of water as a function of the level at the deepest point. Once this survey is completed there is no need to redo it every year.
Different water depths absorb and radiate different amounts of heat throughout the day. Use a quadcopter with thermal imager. Over the course of the day and night, image the fields from fairly high above, take air temp readings at different levels, correlate thermal changes with water depths, integrate for volume.
What are the chances of any survey device getting stuck in the middle of the rice paddy? And if the weight is more than one man can easily pack for 100 meters or so, how are repairs and recharges going to be handled?
This is farm machiery and will need to be fully capable of field repairs and troubleshooting. And don't expect it to not be muddy and wet... expect the unexpected.
I suspect a floating support vehicle would be nearly a necessity to carry tools and new batteries.. maybe to pull the machine out if necessary.
Why would you build anything large and heavy for this? The depth is only 4-7 cm so a small robot in the 5-10 lb range is all that is needed. Personally I would start building something like that with what I already had available, starting with a cordless drill battery and plastic wheels from a grocery cart. I also liked SLRM's suggestion of using an accelerometer for the depth measurement. Add a styrofoam raft around the robot so it won't disappear down a pot hole and a small solar panel and you are good to go.
There must be someone in the Department of Argriculture that has looked at this issue before.
You'd think, but I haven't found it. The FAO has detailed guides on irrigation, complete with percolation and transevaporation formulas. But the depth guidelines seem to be rule of thumb where volume and depth are concerned.
Rice growing in California is very interesting. They have eliminated the need for application of chemicals to their rice paddies by not growing rice for the table. I was told that the entire state's crop is sold to Budweiser for making beer as it doesn't have to be cometically white and beautiful
That's news! You certainly wouldn't know it from the CalRice web site. I thought rice was for sake!
Comments
Rather than measure the body of water directly, one could mix a calibrated amount of a non-toxic environmentally neutral dye in with the chemical treatment. Then fly a drone over the rice paddy to determine that the field has the the right color which would indicate the appropriate measure of application.
Boats may also run into issues, the water depth is small to start with and sensitive to displacement effects.
I'm sure that the simplest possible approach would work the best.
-Tor
The biggest problem, however, is all the localized mins and maximums in the mud, and these add up, as has been mentioned earlier. You can assume they average out but I doubt that's a good assumption - they will probably err on the side of going deeper than the apparent average.
Of course, once you get to a certain point with all this technology, it would probably be cheaper to build an army of robots that march through the mud, hunt down and kill off the snails one by one.
Optical calibration can be done in the field with actual samples of water in the particular field inclusive of sediment on the day of the test and chemical application. This doesn't have to have perfectly clear water or a windless day. The morning usually has less wind, so there may be a particulary time of day that is optimal for using this method.
I do like all the wheeled robot suggestions. A few bike wheels and car window motors would work great. A floating "tail" would then easily probe the surface height. Also, you could do some dead-reckoning with the wheels to augment the accuracy of the GPS. And the final product would be a convenient platform for other samplers and sensors. (i.e. soil chemistry, etc)
Of course the high-tech method would be to use a large area 3D scanner to scan the height of the field just before it was flooded.
Spiking the pesticide with a crop-safe dye is also interesting. You could then optically monitor the concentration of the dye for a closed loop pesticide dosing. i.e. hit the field with half the expected dose, check the concentration some hours later, back calculate the water volume in the field, and adjust the final dose to suit. Thinking further, just dosing with a crop-safe dye and measuring the final dye concentration should allow you to measure the water volume in the field. (assuming negligible flow out of the field and uniform mixing...)
Marty
However, the first step has to be something that can be considered a "gold standard" against which others can be measured. I'm liking the wheel robot too, with the sensor on a spoke or linkage to a float or submerged tail. Wheels for hand trucks are 25cm (10") diameter. A tricycle for stability, or just float the trailing end?
The idea of a dye is intriguing, for back-calibration. It brings up the question of how one goes about disseminating the stuff and mixing it into such a large flat sheet of water. Spray? Tractor? I'll have to find out. Water is lost via percolation into the ground and transevaporation into the air, but the dye marker and the pesticide would have the same concentration dynamics. During the application, all intentional inflow and outflow will cease.
Have a laser up on a pole projecting a line running straight across the field in front of a camera. The camera could then be used to measure the distance between the line on the top of the water and the line at the bottom of the water.
The line will of course be effected by the refraction of the water but I'd think this could be corrected for (I''m pretty sure with the geometry shown, the water will cause the two lines to appear closer together).
(Larger original photo may be found here.)
I liked this idea better previous to making the above illustration. I think the distance between the two lines will be hard to measure for all but the shortest distances. The camera's position and orientation would have to be very accurately known for this sort of setup to have a chance of working.
Edit: I've asked Graeme Finch if it's okay to use his photo this way. I haven't heard back from him yet but if he objects, I'll delete it.
-Phil
The Go Kart place has some big-lug tires, pneumatic. I'd been thinking more along the line of a solid rubber wheel with plastic hub as you would find on a cheap hand-truck. But those are flat, not much tread.
The mud layer holds some fraction of the water and for that reason may need to be counted, so the deep tread may be a feature. Can't over-think this!
Speculation about not being muddy is dependent on local soils and local water sources. Soil is generally classified as a mixture of three non-organic elements -- clay, silt, and sand. And then there are the organic features suspended in the water.
I am beginning to think a narrow probing wheel that runs along the bottom, with a float that records the actual water surface would be best.
Why so? The weight of the robot, inclusive of batteries would get the bot resting on the bottom. It is much easier to monitor the water surface than to have a probe reach the actual bottom with any certainty. Mass would be aligned with a low center of gravity and the weight of batteries would be actually working for you.
I'd still rather do it all with a fly-by survey and not have to run machines over the surface. Machines can fail in many ways and have a lot of down time. And there always is a tendency to revert to gasoline driven machines in such a context.
Duane, that is quite a picture in #39, with the laser tower on the side. I too hesitate to use someone's online photo, what constitutes fair use? I figure a link to the source is the least we can do.
I do hesitate about any optical method. Roughness or undulations or water turbidity could kill it.
Now looking at the path that the rice robot, aka paddy wagon, would have to follow. The perimeter alone is around 800m, and a jaunt to the center brings it up to near 1km. What would be a reasonable speed? Let's suppose 0.25m/s. That means somewhat over an hour to run the course. That is acceptable. But is that a reasonable speed? I attached an animated gif of a possible path that would cover even more points and bring the bot back to its starting point.
This is farm machiery and will need to be fully capable of field repairs and troubleshooting. And don't expect it to not be muddy and wet... expect the unexpected.
I suspect a floating support vehicle would be nearly a necessity to carry tools and new batteries.. maybe to pull the machine out if necessary.
I think you'll get a more accurate and easier-to-compute volume estimate with a trajectory that visits regularly-spaced, orthogonally-arrayed sample points, such as a back-and-forth or square-spiral traverse. Can we assume that the bot will be GPS-guided?
-Phil
-Phil
I favor simplicity. I propose a stick that has a weight at one end, a float at the other, and the electronics inside. The water level is gauged by the angle of the stick and measured with a three axis accelerometer. The position of the stick is determined by a GPS in the float.
In order to cover a whole field the stick has a string attached that can pull the unit to the next location. So, the farmer can easily drag the stick across the field. I've also thought it would be useful to shoot the unit from a bow or ballista.
The advantage here is simplicity: no moving parts. You could also make a bunch of these for very little (<$100 parts) and parallelize your field exploration, or leave one behind to monitor water level over time.
Here's a graph of acceleration as the measurement device goes from vertical to horizontal:
It's more than sensitive enough to measure small changes.
That's an interesting idea, Cody. I could see a couple of guys on opposite sides of the pond with a loop of fish line in between. They pull the ducky across or have a motor do it at constant rate, and then move it a few meters over. Like striping a football field. What device did you use to make the graph? The fact that it is sealed and does not need direct water contact is good point.
Loopy, I too have stories about things that go bump in the night or stick in the mud for outdoor and farm installations. Enough to be cautious and skeptical. Just brainstorming. It's too early to think about repairs and maintenance costs on the vapor!
Tor, I value your observations about the rice paddies. In Japan now? You must have a very nice walk to work. One of my favorite scenes in cinema is the rice planting to song at the end of Seven Samarai. The surviving Ronin observe that the farmers are the victors.
There must be someone in the Department of Argriculture that has looked at this issue before.
Rice growing in California is very interesting. They have eliminated the need for application of chemicals to their rice paddies by not growing rice for the table. I was told that the entire state's crop is sold to Budweiser for making beer as it doesn't have to be cometically white and beautiful.
You might argue that they need to apply fertilizer, but the grower I was talking to would take all the rice stalks to Petaluma and use them for bedding in his chicken farms. And then the bedding with chicken poop was later plowed back into the fields. By producing rice and chickens and selling for beer, the rice growers avoid a big chunk of costs related to farm chemicals.
Taiwan grows rice. In fact it grew about 40% of Japan's rice during WWII. And it has the snail problem everywhere. Eradication has been a huge failure. I am not sure what they do to mitigate. Thailand is a world leading exporter and rice producer. It would be worthwhile to find out how they manage the snail problem.
Japan's rice farming is heavily protected and subsidized by the Japanese government. I am not sure that anything they do reflects economically compeitive reality.
_________________
Tracy,
In a good 3D CAD program like Rhino, you can import a "point cloud," convert it to a surface or mesh, then have the program compute the resulting volume -- all without having to write a line of code.
-Phil
Or, if you have an entire week, you could do it with a Peano curve:
-Phil
I wonder about the affects of wind on 3" of water. The windward side of the a paddy might only be 1" deep while the other side is 5" deep. The hazard is that you start measuring without any wind, and then halfway through the wind picks up and distorts the results.
It just might be the best alternative to rid the area of snails would be a few hundred hungry ducks marched through each paddy at the right time. No chemicals at all. And more revenue from all that duck meat and those duck eggs.
-Tor
EDIT: If the rice paddies aren't level then take a series of aerial pictures with various amounts of dry land exposed, and measure the water depth at the deepest point. Determine the amount of covered area for each depth measurement and you can then compute the volume of water as a function of the level at the deepest point. Once this survey is completed there is no need to redo it every year.
Hey, I'm thinking here.... I'm thinking...
Why would you build anything large and heavy for this? The depth is only 4-7 cm so a small robot in the 5-10 lb range is all that is needed. Personally I would start building something like that with what I already had available, starting with a cordless drill battery and plastic wheels from a grocery cart. I also liked SLRM's suggestion of using an accelerometer for the depth measurement. Add a styrofoam raft around the robot so it won't disappear down a pot hole and a small solar panel and you are good to go.
You'd think, but I haven't found it. The FAO has detailed guides on irrigation, complete with percolation and transevaporation formulas. But the depth guidelines seem to be rule of thumb where volume and depth are concerned.
That's news! You certainly wouldn't know it from the CalRice web site. I thought rice was for sake!