Why NDVI camera when drone deploy can convert?

[rock0305]

What would the benefit of having an NDVI camera be when drone deploy and possibly similar software can convert regular photographs to NDVI? I just don’t understand the extra expense of the camera and having more to lose if the drone crashes or flys off when I can convert images anyways. Am I missing something? Quality?

 

[Peterthedroneguy]

The "conversion" that drone deploy offers isn't real NDVI. Check out this short paper on the problem. If you don't like clicking links I just googled "NDVI white paper ryan nelson"

https://sentera.com/wp-content/uplo...aper: NDVI Filtering Blog&utm_source=PDF File

 

[rock0305]

Interesting read. Thank you for the infomation!

 

[stoltzk]

So it sounds like lots of experience or ag. education would be needed to be able to do these type of inspections?

 

[Peterthedroneguy]

To do them properly and justify the high prices I think that's true. However with aps like AgVault analytics is quickly evolving into a computer skill set so users shouldn't be too worried about getting overwhelmed by higher levels of scientific analysis.

 

[Niggs]

So it sounds like lots of experience or ag. education would be needed to be able to do these type of inspections?

I would say that you would need a really good understanding of NDVI and how that affects crop health/growth if you were going to provide benefit and advice to the farmer but really, that is an agronomist's job. As long as you invested in a proper true NDVI sensor, you could provide the raw survey NDVI data to the agronomist to then advise the farmer.

Many farmers don't actually understand NDVI, or should I say, only forward thinking ones do, but even then, there is often a slight lack of knowledge in what the NDVI data actually means.

It would be too much of a ask for a pilot without NDVI knowledge to then start providing more than just the raw NDVI data to farmers or agronomists. If you were wanting to get into this type of survey, your customer should be the agronomist rather than a farmer.

 

[FCEAgtech]

I would say that you would need a really good understanding of NDVI and how that affects crop health/growth if you were going to provide benefit and advice to the farmer but really, that is an agronomist's job. As long as you invested in a proper true NDVI sensor, you could provide the raw survey NDVI data to the agronomist to then advise the farmer.

Many farmers don't actually understand NDVI, or should I say, only forward thinking ones do, but even then, there is often a slight lack of knowledge in what the NDVI data actually means.

It would be too much of a ask for a pilot without NDVI knowledge to then start providing more than just the raw NDVI data to farmers or agronomists. If you were wanting to get into this type of survey, your customer should be the agronomist rather than a farmer.

EXACTLY!

 

[rock0305]

Thanks for all the replies and thoughts. In my case, I am both the farmer and the drone pilot. I want to figure out what NDVI can do for us but not sure if the right track for me is a true NDVI camera or the DroneDeploy conversions. Any thoughts on that question? Which one would be better for me to figure out, interpret and ultimately make good use of the money and time invested?

 

[FCEAgtech]

Please see my posts in "share some NDVI" True NDVI provides you with an actual scale upon which to work from. NDVI is computed from near infrared imagery, which is a part of the spectrum that your RGB camera does not process. You CAN have a notch filter installed on your RGB camera to capture NIR, but that pretty much negates your RGB image. The value is in being able to capture NIR and RGB in the same flight. In other words, if you are using a P4P, you would have excellent RGB imagery along with your NDVI provided you added that sensor

 

[FCEAgtech]

So you would see an issue in your NDVI imagery (which can be quicktiled and immediately available with a relatively less expensive alternative to drone deploy) and you could instantly zoom into the RGB image in the same software to triage the issue. It will take flights over a number of acres to develop correlations, and everything requires ground truthing, but true NDVI provides the most actionable data. So, I will yet again plug Sentera as the best and most economical path to what you are trying to accomplish. Sentera and a trusted agronomist.

 

[rock0305]

So you would see an issue in your NDVI imagery (which can be quicktiled and immediately available with a relatively less expensive alternative to drone deploy) and you could instantly zoom into the RGB image in the same software to triage the issue. It will take flights over a number of acres to develop correlations, and everything requires ground truthing, but true NDVI provides the most actionable data. So, I will yet again plug Sentera as the best and most economical path to what you are trying to accomplish. Sentera and a trusted agronomist.

Thank you very much for the insight and I will check out that thread and see if I can upload some of the simple stuff I have done. Any other thoughts on the subject will be greatly appreciated as well.

 

[Arbutus]

I would say that you would need a really good understanding of NDVI and how that affects crop health/growth if you were going to provide benefit and advice to the farmer but really, that is an agronomist's job. As long as you invested in a proper true NDVI sensor, you could provide the raw survey NDVI data to the agronomist to then advise the farmer.

Many farmers don't actually understand NDVI, or should I say, only forward thinking ones do, but even then, there is often a slight lack of knowledge in what the NDVI data actually means.

It would be too much of a ask for a pilot without NDVI knowledge to then start providing more than just the raw NDVI data to farmers or agronomists. If you were wanting to get into this type of survey, your customer should be the agronomist rather than a farmer.

Absolutely! You don't want to be in a position of providing professional growing advice if you are unqualified. A grower spends thousands on specific inoculants, pesticides, fertilizers and the wrong advice can be disastrous. Work with an agronomist!

 

[Ivanhoe]

The so-called high precision multispectral camera manufacturers argue that a more precise sensor, narrowband filters with an incident light sensor can provide valuable photometric measurements for NDVI calculations. It is unquestionable that a modified consumer camera is not good for photometric measurements. I think this is not an objective of a farmer for vegetation mapping and not even for NDVI analysis. The objective of the farmers is to have information on their fields for the lowest cost without technical understanding.

The advantages of precision cameras and their photometric datasets are valid in lab environment. The photometric reflectance of a field is affected by many environmental factors (position of the sun -> shadows; strength or direction of the wind -> position of the leaves) which can not be taken into account and which make the precise photometric values almost useless. If we want to compare 2 precision imagesets of the same field made in different time, the results won't be as precise as it is said in the promotions due to the affecting factors. The results are almost as relative as if the images would be made by a converted consumer camera. But the latter one is much cheaper and usually have a much higher resolution.

You can try it for yourselves with any NGB (NIR+green+blue) imagesets here for free:
NDVI Camera | DJI NDVI upgrade | AgroCam | NDVI image processing

An NDVI analysis made by a converted camera can also provide valuable information on the vegetation status of a field.

 

[Arbutus]

@Ivanhoe Unfortunately "converted" rgb images are of little to no value when it comes to actual crop health measurement. RGB cameras produce images which can be visually interpreted of course and there is some value in that, but the plant health is indicated by a very specific spectral range (Near Infra Red and Red Edge) which is the range associated with the most effective conversion of light energy into plant material - i.e. a healthy, growing plant. The NIR or NRE information is not recorded by an RGB sensor.

False coloring and filtering can reproduce similar images to those generated by a 4 band system. Such images are specific to that lighting, azimuth and even temperature environment and would need to be reanalysed for each flight if the lighting, etc. was different. In other words trying to interpret crop health exclusively from an RGB image is vague at best and misleading at worst. Most commercial software and sensor systems are designed to account for variations in the ambient conditions. That is one of the reasons they cost thousands of dollars more than consumer cameras.

 

[rock0305]

Absolutely! You don't want to be in a position of providing professional growing advice if you are unqualified. A grower spends thousands on specific inoculants, pesticides, fertilizers and the wrong advice can be disastrous. Work with an agronomist!

Thanks, I am definitely unqualified but not offering advice to anyone....just using on the farm for myself.

 

[rock0305]

The so-called high precision multispectral camera manufacturers argue that a more precise sensor, narrowband filters with an incident light sensor can provide valuable photometric measurements for NDVI calculations. It is unquestionable that a modified consumer camera is not good for photometric measurements. I think this is not an objective of a farmer for vegetation mapping and not even for NDVI analysis. The objective of the farmers is to have information on their fields for the lowest cost without technical understanding.

The advantages of precision cameras and their photometric datasets are valid in lab environment. The photometric reflectance of a field is affected by many environmental factors (position of the sun -> shadows; strength or direction of the wind -> position of the leaves) which can not be taken into account and which make the precise photometric values almost useless. If we want to compare 2 precision imagesets of the same field made in different time, the results won't be as precise as it is said in the promotions due to the affecting factors. The results are almost as relative as if the images would be made by a converted consumer camera. But the latter one is much cheaper and usually have a much higher resolution.

You can try it for yourselves with any NGB (NIR+green+blue) imagesets here for free:
NDVI Camera | DJI NDVI upgrade | AgroCam | NDVI image processing

An NDVI analysis made by a converted camera can also provide valuable information on the vegetation status of a field.

Thanks for the link. I will check it out!

 

[rock0305]

@Ivanhoe Unfortunately "converted" rgb images are of little to no value when it comes to actual crop health measurement. RGB cameras produce images which can be visually interpreted of course and there is some value in that, but the plant health is indicated by a very specific spectral range (Near Infra Red and Red Edge) which is the range associated with the most effective conversion of light energy into plant material - i.e. a healthy, growing plant. The NIR or NRE information is not recorded by an RGB sensor.

False coloring and filtering can reproduce similar images to those generated by a 4 band system. Such images are specific to that lighting, azimuth and even temperature environment and would need to be reanalysed for each flight if the lighting, etc. was different. In other words trying to interpret crop health exclusively from an RGB image is vague at best and misleading at worst. Most commercial software and sensor systems are designed to account for variations in the ambient conditions. That is one of the reasons they cost thousands of dollars more than consumer cameras.

Having never claimed to be the sharpest tool in the shed, this is a concise, clear explanation that I can understand. Thank you very much!

 

[Ivanhoe]

Arbutus, you are right that 'converted' RGB images have no value in crop health analysis, but you misunderstood something. I wrote about converted consumer cameras (with proper filters and lens) which are able to record true NIR images (without photometric precision) and can be used for NDVI analysis as well as precision multispectral cameras. I wanted to highlight that these cameras can be cheap but useful alternatives to precision multispectral cameras.

@Ivanhoe Unfortunately "converted" rgb images are of little to no value when it comes to actual crop health measurement. RGB cameras produce images which can be visually interpreted of course and there is some value in that, but the plant health is indicated by a very specific spectral range (Near Infra Red and Red Edge) which is the range associated with the most effective conversion of light energy into plant material - i.e. a healthy, growing plant. The NIR or NRE information is not recorded by an RGB sensor.

False coloring and filtering can reproduce similar images to those generated by a 4 band system. Such images are specific to that lighting, azimuth and even temperature environment and would need to be reanalysed for each flight if the lighting, etc. was different. In other words trying to interpret crop health exclusively from an RGB image is vague at best and misleading at worst. Most commercial software and sensor systems are designed to account for variations in the ambient conditions. That is one of the reasons they cost thousands of dollars more than consumer cameras.

 

[Peterthedroneguy]

The so-called high precision multispectral camera manufacturers argue that a more precise sensor, narrowband filters with an incident light sensor can provide valuable photometric measurements for NDVI calculations. It is unquestionable that a modified consumer camera is not good for photometric measurements. I think this is not an objective of a farmer for vegetation mapping and not even for NDVI analysis. The objective of the farmers is to have information on their fields for the lowest cost without technical understanding.

The advantages of precision cameras and their photometric datasets are valid in lab environment. The photometric reflectance of a field is affected by many environmental factors (position of the sun -> shadows; strength or direction of the wind -> position of the leaves) which can not be taken into account and which make the precise photometric values almost useless. If we want to compare 2 precision imagesets of the same field made in different time, the results won't be as precise as it is said in the promotions due to the affecting factors. The results are almost as relative as if the images would be made by a converted consumer camera. But the latter one is much cheaper and usually have a much higher resolution.

You can try it for yourselves with any NGB (NIR+green+blue) imagesets here for free:
NDVI Camera | DJI NDVI upgrade | AgroCam | NDVI image processing

An NDVI analysis made by a converted camera can also provide valuable information on the vegetation status of a field.

Ivan if your point is that there is useful data outside true NDVI that is true. I think for a lot of farmers and agronomist having an RGB image of the field is a huge step forward. But you see this is a symptom of how little data the farmers and agronomist are working with. As we push forward as an industry the market is demanding more and more food out of less and less space. This pressure is the driving force for all this innovation and its my opinion that it will very quickly drive the need past simple RGB data. Precision NDVI is being used already to identify Nitrogen deficiency in fields AND to correlate that with a needed nitrogen amount. Things like weed detection, fertilizer perceptions, and early stress detection that can be done by a computer will quickly obsolete any value that RGB only data or "fake NDVI" provides.

Today I think you have an ok argument but with how fast things are moving tomorrow you may not. Drones systems with enough batteries can run 3-6 thousand. I'm not sure why a 2000 dollar sensor is making or breaking the bank of any one who is serious about being relevant in the next 5-10 years.

 

[Wayne1949]

I would say that you would need a really good understanding of NDVI and how that affects crop health/growth if you were going to provide benefit and advice to the farmer but really, that is an agronomist's job. As long as you invested in a proper true NDVI sensor, you could provide the raw survey NDVI data to the agronomist to then advise the farmer.

Many farmers don't actually understand NDVI, or should I say, only forward thinking ones do, but even then, there is often a slight lack of knowledge in what the NDVI data actually means.

It would be too much of a ask for a pilot without NDVI knowledge to then start providing more than just the raw NDVI data to farmers or agronomists. If you were wanting to get into this type of survey, your customer should be the agronomist rather than a farmer.

I think the main value is to identify areas for the agronomist to do ground truthing so the agronomist can make best use of his time.