PLS-DA model for accurate identification of Chinese cabbage leaf color based on multispectral imaging

Ziwei Xie; Jinghui Yan; Hao Liang; Xiaonan Yue; Xiangjie Su; Huixin Wei; Yin Lu; Xiaofei Fan; Wei Ma; Xiaomeng Zhang; Xiaoxue Sun; Dongfang Zhang; Jingrui Li; Jianjun Zhao; Mengyang Liu; Ziwei Xie; Jinghui Yan; Hao Liang; Xiaonan Yue; Xiangjie Su; Huixin Wei; Yin Lu; Xiaofei Fan; Wei Ma; Xiaomeng Zhang; Xiaoxue Sun; Dongfang Zhang; Jingrui Li; Jianjun Zhao; Mengyang Liu

doi:10.48130/VR-2023-0025

Figures (5) Tables (2)

Figure 1.
Establishment of PLS-DA model for accurate identification of Chinese cabbage leaf color based on multispectral imaging.
Figure 2.
Visual assessment of leaf color phenotype. (a) The leaf color phenotypes of nine homozygous high-generation inbred lines and their F₁ generation. The figure references the Grass green, Emerald green, May green, and Yellow green in 'The Royal Horticultural Society's Color Chart' as color comparisons. Representative materials of (b) 'F₂-442' and (c) 'F₂-449' in F₂ population of Chinese cabbage. (d) Visual inspection on the difference of leaf color of population. (e) Color difference classification of samples. Scale bars = 10 cm.
Figure 3.
Comparison of leaf color identification accuracy of Chinese cabbage based on K-means. Spatial distribution of (a) RGB parameters and (b) HSL parameter identification results. (c) K-means clustering recognition rate of RGB parameters, HSL parameters and multispectral data.
Figure 4.
PLS-DA model parameters. (a) Model fitting index for each sample type on image RGB, HSL color parameters, multispectral data and data fusion. (b) Overall recognition rate of RGB parameters, HSL parameters, multispectral data and data fusion.
Figure 5.
Display the distribution map of SNP index on chromosomes using Gprime's computational model. The Gprime analysis results of BSA sequencing for identifying 'F₂-449' population leaf color classification results based on (a) visual inspection and (b) PLS-DA fusion model (b).

Principal divisor		RGB				HSL			Multispectral					RGB + HSL+ Multispectral
Principal divisor		1	2	3	Count	1	2	Count	1	2	3	4	Count	1	2	3	4	5	6	7	Count
R²X		0.897	0.094	0.008	1	0.751	0.242	0.993	0.659	0.234	0.08	0.009	0.982	0.691	0.191	0.065	0.035	0.005	0.004	0.004	0.995
R²Y		0.236	0.028	0.074	0.339	0.236	0.108	0.344	0.236	0.196	0.024	0.027	0.483	0.238	0.198	0.042	0.025	0.537	0.079	0.047	0.663
Q²		0.231	0.005	0.083	0.299	0.233	0.133	0.335	0.231	0.248	0.009	0.003	0.429	0.233	0.252	0.042	0.023	0.05	0.062	0.087	0.546
Root-mean- square error (RMSE)	A	0.216	0.212	0.21	0.213	0.241	0.191	0.216	0.233	0.203	0.199	0.202	0.209	0.228	0.199	0.199	0.2	0.216	0.215	0.2	0.208
	B	0.376	0.379	0.353	0.369	0.391	0.335	0.363	0.386	0.307	0.312	0.31	0.329	0.385	0.305	0.307	0.304	0.31	0.318	0.323	0.322
	C	0.408	0.414	0.41	0.411	0.405	0.398	0.402	0.403	0.352	0.359	0.361	0.369	0.405	0.354	0.36	0.361	0.376	0.384	0.381	0.374
	D	0.333	0.329	0.276	0.313	0.298	0.272	0.285	0.32	0.286	0.27	0.26	0.284	0.195	0.224	0.245	0.236	0.245	0.277	0.32	0.249
	E	0.39	0.381	0.385	0.385	0.389	0.39	0.39	0.387	0.351	0.248	0.347	0.333	0.388	0.355	0.343	0.334	0.348	0.333	0.307	0.344

Table 1.

Accuracy of PLS-DA model based on image RGB color parameters, HSL color parameters, multispectral data, and data fusion.

Leaf color recognition method Chromosome Position SNV number

Manual visual inspection Chr 9 49732120-60494964 3450
Data fusion-based PLS-DA leaf color identification Chr 9 53002792-53999899 1539

Table 2.
Candidate region statistics.

Leaf color recognition method	Chromosome	Position	SNV number
Manual visual inspection	Chr 9	49732120-60494964	3450
Data fusion-based PLS-DA leaf color identification	Chr 9	53002792-53999899	1539