Developed and Written by
Dr T.O Amusa
The Trees Identification mobile application is developed to bridge the gap and aid easy identification of trees, especially trees that are native to the Sudan Savanna region of Nigeria. Thorough research was carried out to identify the species, which were then turned into a database to provide more accuracy than an existing tree and plant identification software. You will understand extensively as we go into the details of this doc.
This documentation is written and will be regularly updated to provide details of this software, its stability, framework, uses and future development and improvement. The documentation will discuss the technology breakdown and functionality of the software to aid development continuity.
1.1 STATEMENT OF PROBLEM
The task of identifying plant species with sole reliance on dead specimens that have been pressed, dried, or pricked is becoming more and more tasking by the day, and the pressure constantly being mounted on the fragile specimens has continued to increase at times resulting in some of these age-long conserved specimen being destroyed. There is also the problem of using mobile application software with little or no accuracy or prior research, which is somehow misleading.
The recent adoption of mobile learning applications to aid these identifications has become valuable, Although, many times, the accuracy level of these mobile learning applications has been questionable, and the fact that this software requires internet and mobile data, which may technically be unavailable in remote locations where plant identifications are primarily carried out. That defeats the point of precise identification and inspires thorough research on the accuracy level of this digital software in conducting a complete taxonomy of trees and plant species.
1.2 SOFTWARE FUNCTIONALITY BREAKDOWN (EDUCATION)
Until the app is ready for expansion and improvement, we shall keep the scope of our research to the University of Ilorin Campus while tree species are obtained, pricked, dried and processed in a laboratory. You will understand why this is important as you continue to digest this detailed documentation.
The aim of this software is to use a unique system of scientific identification that serves as an all-rounder and can be proven useful in educational teaching, studying and routine identifications of trees and plants. It is also meant to provide more accuracy than most of the existing software; therefore, research is required to ascertain and identify existing plants from the selected locality.
First, training sets were created by carefully evaluating and picking the sixteen most distinguishing properties of tree species that may be utilized to identify them most simply by themselves. This laid the groundwork for the creation of binary codes, which are used to identify the tree species. The presence or absence of the sixteen morphological features used to categorize each species results in a unique code for that species. Binary code is formed by assigning a “1” to a character’s presence and a “0” to its absence. Secondary binary codes are created utilizing other morphological traits that were not initially considered for selection in order to uniquely identify species where the same binary code has been generated for the identification of more than one species.
- The IDENTIFICATION system keys
This key is adapted from the main keys of the proposed Flora of Nigeria. In that work similar keys will be provided for the flowers and fruits, but for trees, these are seldom readily available and often difficult to observe. The key is based on the following 16 vegetative characters:
- Other exudates
- Simple leaves
- 1-pinnate leaves
- Opposite leaves or leaflets
- Leaves or leaflets widest above middle
- Leaves or leaflets widest below middle
- Blade of leaf that is very narrow
- Blade of leaf that is very broad
- leaf or leaflet whose margin is entire
- leaves or leaflets whose blades are over 30 cm
- leaves or leaflets whose blades are under 2.5 cm
- leaves or leaflets whose stalks are over 5 cm
- leaves or leaflets whose stalks are under 1.75 cm
- Spines or thorns (in this book, spines are regarded as being straight and thorns curved)
maybe present on the stems or branches as in Fagara or Bombax, at the base of the leaves as
in Oncoba or Acacia, on the leaf stalks as in Erythrina or on the margin of the leaves as in Rinoreailicifolia.
- Latex is defined as being opaque and milky-white, as in Euphorbia, Colourless juices are excluded, and so are coloured exudates. The latex may be observed by slashing the bole or, if living material is available, by cutting a branchlet or tearing a leaf across. Latex-producing trees usually show it in the slash, but occasionally (Rauvolfia caffra is an example) the latex appears only in the leaves and young shoots.
- ‘Other exudate’ includes conspicuously red, yellow or orange exudates (whether or not
they are resinous, gummy, or sticky as well) together with all resinous, gummy, or sticky exudates. White latex (whether sticky or not) and watery non-sticky liquids which are colourless (or of some other colour than the above) are excluded. Many exudates change colour: the colour of the exudate that first appears after the tree has been slashed is the one that should be recorded; it must be remembered that some exudates are slow to appear.
- Simple leaves includes deeply divided leaves as in Triplochiton or the Pawpaw.
5. Multi-pinnate leaves.
- In addition to a pair of leaves on opposite sides of a branchlet, opposite’ includes simple leaves arranged in a whorl of 3 or more, as in Alstona. [he leaves must be clearly opposite or
whorled and not merely in a closely arranged spiral as in Terminalia. In compound leaves, the leaflets, and not the whole leaf, are to be considered. In some families, Meliacaea
and Sapindacene in particular, the character is rather variable: whether the leaflets are opposite or alternate should be decided from inspection of the tree as a whole rather than from an individual leaf.
The key, however, provides for either answer where the variability is a normal character of the species. In digitate compound leaves the leaflets are always regarded as being opposite.
- The middle of a leaf or leaflet is half-way between the apex and the point of attachment to the stalk. The character should be obvious for the majority of the leaves on the tree: the rule is to put 0 in doubtful cases. In compound leaves only, the longest leaflet should be considered.
- This character defines very narrow leaves or leaflets. In the case of tiny leaflets, as in Acacia or Piptadeniastrum, the measurement is a matter of difficulty and the key provides for either alternative.
- In assessing the ratio of length to breadth the whole blade is considered (not merely
from the tip to the point of attachment as in the 7th and 8th characters).
- Entire refers to the sides of the blade of a leaf, whether or not the apex or base is indented. A wavy margin is entire, but where there are, definite lobes the margin is not entire. When there are minute teeth, as in Ziziphus, or glandular teeth, as in Ricinodendron, the margin is not entire.
- The 12th character is positive if some of the leaves or leaflets are over 1 ft. long though most of them may be less than this. In assessing this character and the next the aim should be to collect the crown leaves of mature trees since saplings and sucker shoots bear leaves often differing in shape and usually much larger than adult ones. If this is not possible and it appears
- In this case none of the leaves or leaflets should be over 1 in. long.
- Only some of the stalks require to be over 2 in. long. Only very rarely are leaflet-stalks in compound leaves as long as this.
- In dealing with compound leaves it should be remembered that it is the leaflet-stalk that is to be considered. Thus, in Bombar the leaves have long stalks but O would have to be put for this character since the leaflets are practically stalkless.
- This character defines leaves with palmate main nerves as in Triplochiton. Leaves with 3 main nerves from the base, as in Celtis and some species of Memecylon are also included; species of Albizia.
A user supplies answers in Yes/No format to questions that the system has been designed to ask based on the selected morphological characteristics on which the training sets have been created.
(a). The system then generates a binary code by converting every “Yes” answer to a 1 and every “No” answer to a 0. The length of this code is usually 16 representing the 16 morphological characters being used in stated keys for identification.
(b). The system searches its inference-rules-base for a satisfactory rule using the binary code generated; if found, the species linked to that code is returned as the name of the species to be identified
The result that will be returned from the users’ attempt using the following examples is a plant species called Acacia gourmaensis with details of the tree and images. For accuracy.
- Step 1: The user opens the app and types in their name and email
- Step 2: The Questions GUI Pops up, and the user begins to answer the questions
- Step 3: At the end of selecting true or false for the 16 questions in the software, a request is sent to the database to check if there’s an answer available for user selection
- Step 4: If an answer has been entered for the binary code generated from the user’s selection, the answer is returned, and the software is successfully able to identify the tree for the user
- Step 5: If no answer is selected, the software returns, ” No option Available” Send the tree image to admin for identification. This may take 7 business days; if the user completes the form and sends the image they are trying to identify to an admin, the admin will receive an email and can help the user from there.
App GUI for the admin area
- ADMIN AREA
The admin area is created to bridge the gap of managing live codes, We understand that managers do not have to be someone vast in any of the core programming languages or framework used to build the software hence a user interface is provided to manage the software. This area is not accessible to general application users but just the administrator. An Administrator will be able to perform the following functions
- Upload, Edit and Delete the 16 system dependent questions
- Search for code binary to update a newly identified tree or an already updated tree that requires a second look or review
- Edit the Botanical and Common Name of species
- Upload an image of already identified trees to provide more assurance on the correctness and accuracy of an identified tree
- Review history of users who recently used the software
- Provide additional explanation on questions asked within the software
- USER AREA
The user area is the interactive interface where user connect with the software and provide answers to the 16 system generated questions to aid their identifications. It has the following
- Collation of users attempted questions
- All questions are required to generate correct result
- Table displaying identified trees and results
- More information and explanations on questions where scientific term is used
1.3 SOFTWARE FUNCTIONALITY BREAKDOWN (TECHNOLOGY)
Uploading… (Please check back later)