analyse_object error - Githubissues

gdauby commented 1 year ago

Hello

I am testing this very promising package.

I have an error with one photo attached, using the function analyze_objects

library(pliman)

leaves <- image_import("IMG-2581.jpg", plot = TRUE)

count <- analyze_objects(leaves,
                         marker = "id")

Error in EBImage::distmap(img2) : 'x' shouldn't contain any NAs

Anything I am doing wrong ?

Thanks

TiagoOlivoto commented 1 year ago

Hi @gdauby , I couldn't reproduce the error on my side. Please, verify if you are using the dev version of pliman (make sure to install pliman from github to have access to the last news and bug corrections (I'm planning a major release on CRAN in a few weeks). To install from github, use remotes::install_github("TiagoOlivoto/pliman")

Here're some tips to analyze such image.

library(pliman)
#> |==========================================================|
#> | Tools for Plant Image Analysis (pliman 1.2.0)            |
#> | Author: Tiago Olivoto                                    |
#> | Type 'citation('pliman')' to know how to cite pliman     |
#> | Type 'vignette('pliman_start')' for a short tutorial     |
#> | Visit 'http://bit.ly/pkg_pliman' for a complete tutorial |
#> |==========================================================|
setwd("E:/Downloads")
# set_wd_here() # set the working directory to the path of file
img <- image_import("img.jpg") 

# first of all, you need to choose an index to segment
# 1. leaves and reference from the background (keep as foreground)
# 2. leaves from reference (keep leaves)

# the following, will perform the segmentation with all available indexes
# image_segment_iter(img, nseg = 2)

# after some tests, the two indexes are B and R-G
image_segment_iter(img, index = c("B", "R-G"))
#>      image   pixels   percent
#> 1 original 22943448 100.00000
#> 2     seg1  8181438  35.65915
#> 3     seg2  8101020  99.01707

# note that fill_hull = TRUE may help you to fill holes
# you can also test the 'filter' argument.. this will remove possible noises

# Now, you can use analyze_objects, with the the following arguments
# the reference object will be colored in red (this will help you to see if the segmentation was performed correctly)
res <- analyze_objects(img, 
                       reference = TRUE, # indicates that there is a reference in the image
                       back_fore_index = "B", # index to segment reference + leaves from background
                       fore_ref_index = "R-G", # index to segment reference and leaves
                       watershed = FALSE, # set to TRUE if leaves are touching each other
                       # object_size = "large", # this will control how watershed works.. try changing to small
                       marker = "area",
                       filter = 3, # remove noise
                       fill_hull = TRUE, # if FALSE, will result in several 'holes' in the leaves
                       reference_area = 4) # I assumed that the reference has 4 cm2, change according

res$results
#>   id         x        y      area   area_ch perimeter radius_mean radius_min
#> 3  3  381.7126 1348.023 109.80755 111.68415  44.81007    6.119369   4.124521
#> 4  4 1539.8681 1469.431 119.17742 121.80397  48.61357    6.424274   4.312093
#> 7  7 1593.0025 3085.195 111.60273 112.41278  45.41554    6.188672   4.312488
#> 8  8  372.5171 3127.172  67.94885  68.79991  37.53980    4.978166   2.893091
#>   radius_max radius_sd diam_mean diam_min diam_max major_axis minor_axis
#> 3   9.126145  1.495792 12.238738 8.249043 18.25229   16.67364   8.418830
#> 4   9.912393  1.527892 12.848548 8.624187 19.82479   17.26657   8.825506
#> 7   9.493167  1.583863 12.377345 8.624976 18.98633   17.20527   8.303927
#> 8   8.104797  1.575048  9.956331 5.786181 16.20959   14.85112   5.850375
#>     length    width radius_ratio eccentricity     theta  solidity convexity
#> 3 17.98685 8.626488     2.212656    0.3471054  1.561144 0.9831973 0.8981382
#> 4 18.61489 9.100198     2.298743    0.3307280  1.445139 0.9784362 0.8417816
#> 7 18.87298 8.751091     2.201320    0.3266697 -1.538252 0.9927940 0.9189865
#> 8 15.98549 5.975648     2.801432    0.2299912  1.416736 0.9876299 0.9147805
#>   elongation circularity circularity_haralick circularity_norm   coverage
#> 3  0.5204003    18.28601             4.091057         1.458265 0.09628134
#> 4  0.5111334    19.82992             4.204665         1.581383 0.10449702
#> 7  0.5363163    18.48137             3.907329         1.473814 0.09785539
#> 8  0.6261830    20.73967             3.160644         1.655233 0.05957884
#>          asm      con       cor       var       idm       sav       sva
#> 3 0.04002358 2.705853 0.7732885  6.967612 0.6046174  8.175257  70.01186
#> 4 0.02118343 2.881822 0.9005808 15.493289 0.5958295 16.751592 289.50067
#> 7 0.03846488 1.495121 0.9385237 13.160129 0.6938088 16.850686 287.62320
#> 8 0.03103248 1.506769 0.9401819 13.594603 0.7055227 16.120046 265.57900
#>        sen      ent      dva       den       f12       f13
#> 3 1.186175 1.631566 2.705853 0.5830037 0.2156136 0.5714213
#> 4 1.441160 1.917670 2.881822 0.5930027 0.3361421 0.7345269
#> 7 1.343339 1.678359 1.495121 0.4842656 0.4114297 0.7621291
#> 8 1.399063 1.724419 1.506769 0.4773082 0.4498587 0.7953645

^{Created on 2023-02-14 with reprex v2.0.2}

Please, let me know if it worked!

gdauby commented 1 year ago

Thanks for the tips. It helped a lot.

TiagoOlivoto / pliman

analyse_object error #2