Annotating observations with multiple photos

I thought I understood how to annotate plant phenological characteristics but apparently not. Rather than point to a specific example, I’ll concoct the simplest example I can think of.

Suppose we have an observation with two photos. Assume all plants in both photos are the same species. In photo #1, there are flowering plants only, while in photo #2, there are budding plants only. How should the observation be annotated?

Thanks in advance.

If it’s 2 individuals, I think you should ask the observer which the observation is for (or decide on one with the community, as often happens when people are unresponsive…). Each observation is designed for 1 individual at a time, even though this often isn’t strictly adhered to. It’s usually not an issue, except for corner cases like these.

If it’s 1 individual at obviously different times, I think you should determine, in the same way, which one is correct for the obs if possible. I think this will probably only rarely be possible if the observer hasn’t specified or it’s not clearly the wrong season.

If it’s 1 individual with both phenologies, …I should wait for an expert to answer :)

1 Like

I thought you could add multiple annotations.

The annotation is at the observation level, not at the photo level. I have many observations showing 1 individual plant but each photo may show various parts of the plant some flowering, budding, fruiting, multiple, or none of the above.

In your example, if it’s all one plant/clump, it would be annotated as both budding and flowering.

3 Likes

What if it is flowering, but last year’s fruit is still hanging on.

Plants allow multiple annotation, so annotate everything you see.

3 Likes

according to :
https://www.usanpn.org/files/shared/files/Haggerty&Mazer_ThePhenologyHandbook_v3Aug2009.pdf

42 Table 3: Herbaceous plants

Open flowers In at least one location on the plant, an open fresh flower is visible. Flowers are considered “open” when the reproductive parts are visible between unfolded or open flower parts. Do not include spent (wilted) flowers that remain on the plant.

Ripe fruits In at least one location on the plant, a ripe fruit is visible.

3 Likes

My post is valid for plants; other organisms I am less familiar with. Typically, the phenological stage is assessed for the whole plant, not parts of it. So if you look at a tree where not all the flower buds open exactly at the same time, you would assess what percentage is opened. On a time serie where you would be interested in documenting the development of the plant, each time you go back to the field to collect data, those would be considered separate observations from the previous times, even if it is for the same individual. Anyhow, recording it that way would make things a lot simpler for subsequent statistical analysis. Also, I would look for specific plant species what scale the professionals in the field use because it tends to vary, and what I am most familiar with are cultivated species. For cultivated plants, the BBCH is the most universal and a generalisation of the Zadoks scale which is used for wheat. I hope this helps.

1 Like

I think if you can see buds, flowers or fruit in any picture in your observation, you can add the annotation. And there aren’t that many people who add the annotations, so I appreciate those who do.
I personally use the filter when I’m looking observations to compare to what I suspect my own observations might be. A few buds or fruits among the flowers isn’t going to ruin that, as long as I can see the flowers I’m looking for. It still helps filter out all the noise.

4 Likes

I often mark the observation as flowering and budding, sometimes flowering and fruiting, sometimes all three. I just mark what I see, whatever it is. I’m glad we don’t have to choose between these options.

Technically, each observation is supposed to be just one individual plant (which might be in multiple phenological states). Technically, two different plants even if the same species should be separate observations, though if they’re photo’d at the same time and place I’m not going to waste energy being upset about their being treated as one. If they are plants photo’d at different times, they should be posted as different observations. If they’re the same individual posted at different times, it’s really valuable to post them as separate observations and include a statement with the iNaturalist observation of the other observation in each of them.

3 Likes

Thanks everyone. After reviewing your answers, I think I’ve been doing it wrong. I’ve been pretty strict (perhaps overly strict) about applying the “one plant per observation” rule-of-thumb. If it’s obvious there are multiple plants depicted in multiple photos, I either ask for a change or ignore the observation altogether. I should probably just do the latter.

That said, given this fairly relaxed point of view, I don’t see how plant phenological data can ever be utilized by the computer vision algorithm. Perhaps that was never the intention. I’m not sure where I got that notion.

1 Like

[off topic]

Wow!

I’m more flexible than this (e.g., unripe fruits are fruits in my book) but I could live with whatever reasonable guidelines the community might adopt.

1 Like

So, could you provide a working link, please? It doesn’t open even with proxy.

Works for me. Network issues?

2 Likes

This document might help as it illustrates phenological stages for dicots and monocots of both ligneous and herbaceous plants. BBCH Monograph

1 Like

I dunno, it says ERR_EMPTY_RESPONSE, and did the same yesterday.

1 Like

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.