[rejsharp]

GOOD IDEA – could also be done online if we had some search and replace tools (including search for null values).
All databases need housekeeping tools, and we are bound to make mistakes or get smarter ideas.

[rejsharp]

Hi Randhir : There is an Android App called BLE Checker. You can see if your phone supports BT4 or not.

Please note that BLE = Bluetooth Low Energy = Bluetooth Smart = BT 4.0 and newer.

It needs the right hardware, and new phones coming with Android 4.3 and above should have it.
BUT – the BT implementation under Android 4.3 was rubbish. Android 4.4 was better, and Android 5 seems quite good.

Also note that upgrading the OS version on an OLDER phone does NOT bring BTLE as the hardware is not there.

Best regards, Roger

[rejsharp]

Amit, I am not at sure that you are on the right track.

The SCiO scans are influenced by molecules rather than elements.
So the spectra from methyl alcohol CH3-OH will be different from that of ethyl alcohol C2H5-OH.
It does not work on elemental signatures of carbon, hydrogen and oxygen.

If I wanted to make an estimation model to determine how much ethanol was in an unknown mixture,
I would make up standard reference samples going in 5% steps from 0% to 100%

If they were aqueous solutions I would need a tertiary model for my reference samples.

The attributes for each sample would be methanol %, ethanol %, water%, and sampleID

Maybe you can explain further what your objective is.

[rejsharp]

Thanks for that! Do you have a prototype or a finished production item?

[rejsharp]

Hi Philippe,
I scanned all three places cap, underside and stem.
Mostly the scans are very different between the three, even with the same actual mushroom.
Not enough samples scanned yet, but the underside looks promising. I am worried about cap scans because of variability – we will see.

It is extremely hot and dry in Austria now, so no wild mushrooms to be found at the moment.

regards — Roger

[rejsharp]

Yes you are right! With a series of test mixtures as reference, the SCiO should be able to estimate the percentages in a mix.
(Sorry for me misreading your text, and muddying the water with solutions).
Ahhh there is of course room for error with powders of different grain sizes / densities – they will segregate out for fun! The fine stuff ends up on the bottom.

I realise this has drifted a long way from Nikos’ original post theme which was asking about validating sample contributions. This is a critically important subject (Rubbish in guarantees Rubbish out). Consumer Physics did mention this a while ago, and I will raise the subject at the Paris Workshop.

[rejsharp]

Hi Redwingii – you cannot combine dry salt scans in the same model as liquid salt solutions. The molecular state is different so NIR resonance is hugely different.

I have added some scans of dry salt and sugar, and solutions of them into our previous thread.
Roger

[rejsharp]

Hi Redwingii – I think that this belongs here rather than projects.

I think that the salt and sugar scans for estimation need to respect physical state.
Dry salt and Dry sugar will scan very differently to solutions, because the molecular state has changed.

Note 1: The liquids were scanned in a white porcelain cup, I am still not sure about the best way for liquids.
Note 2: Liquids were intended to about 50% by weight, but not all the salt dissolved.

Attachments:

You must be logged in to view attached files.

[rejsharp]

Hi Nikos,

The raw data seems to be a pure measure of light intensity per wavelength, so there is considerable variation depending on ambient lighting, sample distance, sample irregularity, temperature,  SCiO internals etc.

It is possible in the SDK to download these actual data, but no stats are generally available on deviation of either raw or processed scans. The models do give a reliability index.

I see that CP have just added a new processing method, but I have not tried it out yet.

best regards, Roger.

[rejsharp]

Hi Redwingii,
I think that the key to solve this is to dry as completely as possible, but with frequent intermediate weighing and associated scans.
Plotting the weights in a table will let you see when further weight loss has stopped, this will be close to zero % moisture as long as you are not cooking your sample.
– Now you can work back to estimate moisture at the earlier scans.
– Add this moisture content as an attribute measurement of each scan
– Try a Model to see if you get good results
Now try a fresh leaf, see what SCiO, predicts for moisture content, then measure the moisture by drying completely.
– If it is within a couple of percent – you are a winner! If not, more work is required.

Note 1: work on several leaves in parallel, scan each 5* at each stage. Don’t leave too much time between samplings to avoid biochemical changes. (Balance with rate of drying).
Note 2: you will need a very accurate balance, and good lab procedures (always weigh at a fixed temperature, avoid breezes etc).
Note 3: essential oils may evaporate also, keep dehydration temperatures gentle. Maybe estimate their quantity by washing a chopped leaf in ethanol (maybe ether?) dry both the washed leaf, and solvent to get base data. Repeat this for your slowly dried leaf. Remember to wash and dry a blank filter paper for comparison.
Note 4: A sachet of silica gel suspended near your leaf will help dehydration
Note 5: You could also scan a larger number of leaves and dry each completely, but the range of starting values may be too tightly clustered to get a reliable model.

Someone more skilled in lab practices or botany may laugh with scorn! Please feel free to criticise – I am not proud, and always ready to learn.

Roger

[rejsharp]

Yes, I will attend the Paris meeting, afternoon session.

[rejsharp]

Chris, you need more samples, and attributes that are different between samples.
Scan the powder again 3x as a New Sample, then again 3 more, then again 3 more.
Perhaps with different temperatures / lighting conditions.
Then scan a different powder similarly. Now you have something that SCiO can model.
I may have the wrong understanding of what you have done, but if you create a Spices collection, and only have 12 scans of black pepper as one sample, it is impossible for SCiO to create any model. You need multiple samples and multiple materials.
I hope that helps.
Roger

[rejsharp]

Booked for Paris Meet-up July 22nd!

[rejsharp]

Hi Redwingii:
Did you progress any with your scanning aid distance piece?
I see that CP are looking at making one also.
I have tried some photographic black-out cloth, it does work shielding the scan from bright sunlight, but is a bit fiddly as it likes to curl up. Then you need four hands! One for SCiO, one for smartphone, one for sample, and one for the shield.

I am sure that there are enough clever engineers amongst us to create a solution.
Roger

[rejsharp]

Luis, I have some photography blackout cloth (rubberised), I can send you a strip if you are interested.
I used it successfully to shield scans from sunlight. It was a bit fiddly, as the cloth likes to curl up if not restrained.
I am interested to see what the shield looks like that CP mentioned under Scanning Aids.
Roger.

[Author]

Posts