E cognition object based classification would help,

Do you need the individual positions or just a total count? If you are only interested in the total count, a very simple technique of using individual pixel colors as features for a standard size image and using a regression model for count can give you usable results. For identifying individual trees you may dive in some clustering/segmentation methods combined with supervised classification.

However, you need labeled data for using the above models. I would recommend using LiDAR in a limited area to generate labeled data that can then be used to develop predictive models based on satellite imagery.

Article Assessment of Individual Tree Detection and Canopy Cover Est...

Neetesh Sharma, excellent answer. Indeed, count depends on area of research and area of canopy, it clear visible on attached image. No noise produced by another trees

Hi

20.000 Ha is possible with our long range heavy lift drone www.geospace.co.nz

when coupled with our Lidar sensor for data capture and automatic canopy classification software then the tree crown count can be established even when there is such dense growth that no lidar pulses penetrate the canopy to find ground level.

contact me [email protected] to discuss in detail how this can be achieved

this article might help you:

Article A review of methods for automatic individual tree-crown dete...

This PhD thesis might be useful also for rubber plantations: Mahour, M. (2018). Scaling of Remote Sensing Information for Orchard Management. Enschede: University of Twente, Faculty of Geo-Information Science and Earth Observation (ITC). https://doi.org/10.3990/1.9789036546898

Options such as using foliar and/or canopy traits through quantitative remote sensing can help! This can be achieved using, for instance, key traits (biophysical, structure, etc.) specific to rubber trees, and then combined with radiative transfer modelling (RTM), for instance, Discrete Anisotropic Radiative Transfer (DART) (https://dart.omp.eu/#/), attributes of individual trees can be retrieved, and hence, help you to know the number of the trees to which the obtained attributes belong. Sorry that my answer may sound a bit vague, but feel free to check some publications which can be greatly helpful such as "Remotely Sensing the Species of Individual Trees": https://library.itc.utwente.nl/papers_2020/phd/shi.pdf or Article Individual Tree Detection in a Eucalyptus Plantation Using U...

All the best,

Maurice

Geographic Object-Based Image Analysis provides decent solutions to your question.