Multitemporal synthetic aperture radar (SAR) imagery provides continuous, cloud-free observations of crop phenology for agricultural monitoring in cloudy and rainy regions, yet high-accuracy classification typically requires extensive labeled samples. Although self-supervised pre-training mitigates label dependency in remote sensing, existing frameworks primarily target optical imagery or generic spatial SAR scenes, underutilizing temporal backscatter dynamics and polarimetric scattering mechanisms. We propose SAR-TS-BERT, a physics-guided BERT-style self-supervised pre-training framework for SAR time-series crop classification. SAR-TS-BERT integrates backscatter intensities and polarimetric ratio features (VV, VH, and VV/VH) with dual-polarimetric Cloude-Pottier decomposition parameters (H/α/A) into a six-band input representation through two core designs: 1) a physics-guided span masking strategy (SpanPhysics) that adaptively selects contiguous masked spans based on temporal backscatter change rates to prioritize key phenological transitions; and 2) a dual-branch cross-modal fusion encoder that models backscatter and polarimetric decomposition features through separate Transformer branches followed by bidirectional cross-attention fusion. On independent test sets in Estonia (EuroCropsML) and the USA (California CDL), SAR-TS-BERT achieves overall accuracies (OAs) of 96.78% and 87.29%, respectively, and attains a mean monthly OA of 99.37% on Chongming Island, Shanghai. On the Estonia benchmark, pre-training improves OA and Macro F1 by 1.17 and 1.59 percentage points over training from scratch under full supervision, and by 2.58 and 5.19 percentage points using only 1% labeled data. Incorporating SAR scattering physics into temporal self-supervised pre-training effectively enhances crop classification accuracy, parcel-level mapping consistency, and few-shot transferability.
Figures & tables
Fig. 1: Study areas: (a) California, USA (Mediterranean climate); (b) Estonia (Baltic region, cloudy and rainy climate); and (c) Chongming Island, Shanghai (subtropical monsoon climate).
Attribute
Estonia
USA
Chongming
Climate
Cloudy/rainy
Mediterranean
Subtropical monsoon
Spatial extent
National
Central Valley
Island
Granularity
Parcel-level
Pixel-level
Parcel-level
Total samples
5,664 parcels
∼ 28.9M pixels
543 parcels
Crop classes
9
8
11
Time span
2021
2024
2023.12–2024.10
TABLE I: Overview of the Three SAR Time-Series Datasets
Crop Type
Samples
Proportion (%)
Pasture/meadow/grassland
260,846
48.29
Winter wheat
70,395
13.03
Spring barley
45,454
8.41
Legumes harvested green
31,266
5.79
Clover
30,765
5.70
Peas
30,389
5.63
TABLE II: Class Distribution of the Estonia Dataset (Within-Parcel Pixel Samples)
Crop Type
Samples
Proportion (%)
Grapes
9,519,034
32.88
Almonds
5,985,871
20.68
Walnuts
3,599,395
12.43
Corn
3,133,946
10.83
Alfalfa
2,191,909
7.57
Tomatoes
2,187,955
7.56
TABLE III: Class Distribution of the USA Dataset
Category
Samples
Proportion (%)
Rice
181,435
40.15
Wheat
105,514
23.35
Plowed land
69,486
15.38
Rice stubble
31,972
7.07
Woodland
17,880
3.96
Rapeseed
11,919
2.64
TABLE IV: Crop and Land-Cover Class Distribution of the Chongming Island Dataset (Monthly Pixel Samples)
Fig. 2: Overall architecture of SAR-TS-BERT. (a) A six-band SAR time series comprises backscatter features (VV, VH, and VV/VH; blue) and polarimetric decomposition features ( H , α , and A ; green), where image thumbnails and waveforms are illustrative. (b) Contiguous span sampling weights are computed from temporal backscatter change rates, and additive Gaussian noise is injected into all six bands at selected masked timesteps. (c) The two feature groups undergo separate linear projection, sinusoidal positional encoding concatenation, and branch-specific Transformer encoding, followed by bidirectional cross-attention and a fusion Transformer to generate temporal representations. (d) During pre-training, a linear reconstruction head restores the six-band observations and computes mean squared error (MSE) loss on masked timesteps; during fine-tuning, the pre-trained encoder is coupled with temporal mean pooling and a linear classification head trained with cross-entropy loss.
Stage / Component
Parameter
Setting
Pre-training
Epochs
100
Batch size
512
Optimizer
AdamW
Learning rate
1×10−4
Masking ratio
pmask=0.15
Span length
L=3
TABLE V: Implementation Details of SAR-TS-BERT
Method
Params
OA
Macro F1
Weighted F1
Random Forest
-
0.8577
0.7946
0.8429
BiRNN (2-layer)
403K
0.8987
0.8628
0.8976
BiLSTM (2-layer)
537K
0.9023
0.8681
0.9002
SITS-BERT-SAR [ 5 ]
3.20M
0.9501
0.9335
0.9498
PSE-TAE-SAR [ 3 ]
3.46M
0.9427
0.9239
0.9427
ExViT-TS [ 33 ]
3.17M
0.9316
0.9107
0.9316
TABLE VI: Main Quantitative Comparison on the Estonia Dataset
Crop Type
F1 (From Scratch)
F1 (With Pre-training)
Δ F1
Test Samples
Clover
0.9044
0.9366
+3.22%
15,704
Legumes harvested green
0.8895
0.9205
+3.10%
15,516
Oats
0.9293
0.9518
+2.25%
12,255
Pasture/meadow/grassland
0.9710
0.9795
+0.85%
129,742
Peas
0.9607
0.9693
+0.86%
15,233
Spring barley
0.9488
0.9595
+1.07%
23,080
TABLE VII: Per-Class F1 Score Comparison on the Estonia Dataset
Method
Full Parcel OA
Accepted OA
Accepted Macro F1
Coverage
Rejected Parcels
Accepted Wrong Parcels
SAR-TS-BERT
0.9958
0.9982
0.9980
98.69%
37
5
PSE-TAE-SAR
0.9944
0.9964
0.9958
98.38%
46
10
SITS-BERT-SAR
0.9905
0.9946
0.9939
98.09%
54
15
ExViT-TS
0.9901
0.9935
0.9927
98.13%
53
18
Presto-SAR
0.9862
0.9931
0.9923
96.61%
96
19
TSViT-SAR
0.9848
0.9931
0.9921
96.93%
87
19
TABLE VIII: Threshold-Constrained Parcel-Level Mapping Results on the Estonia Test Set (Total Test Parcels: 2,832)
Fig. 3: Comparison of parcel-level crop prediction maps on the Estonia test set.
Fig. 4: Spatial distribution of accepted wrong parcels and uncertain/rejected parcels on the Estonia test set. Red indicates parcels accepted under the voting threshold ( rp≥0.6 ) but misclassified; black indicates low-confidence rejected parcels ( rp<0.6 ).
Method
Params
OA
Macro F1
Weighted F1
Random Forest
-
0.8232
0.7832
0.8191
BiRNN (2-layer)
403K
0.8182
0.7656
0.8157
BiLSTM (2-layer)
537K
0.8200
0.7761
0.8169
SITS-BERT-SAR [ 5 ]
3.20M
0.8689
0.8547
0.8663
ExViT-TS [ 33 ]
3.17M
0.8613
0.8458
0.8606
TSViT-SAR [ 34 ]
4.83M
0.8605
0.8437
0.8595
TABLE IX: Main Quantitative Comparison on the USA Dataset
Crop Type
F1 (From Scratch)
F1 (With Pre-training)
Δ F1
Test Samples
Corn
0.9358
0.9463
+1.05%
2,350,136
Winter Wheat
0.9558
0.9660
+1.02%
1,589,918
Alfalfa
0.8542
0.8836
+2.94%
1,644,062
Tomatoes
0.9446
0.9531
+0.85%
1,642,029
Grapes
0.8406
0.8635
+2.29%
7,139,181
Almonds
0.7667
0.8014
+3.47%
4,489,549
TABLE X: Per-Class F1 Score Comparison on the USA Dataset
Method
OA
Macro F1
Weighted F1
SAR-TS-BERT
0.8766
0.8663
0.8760
SITS-BERT-SAR
0.8569
0.8399
0.8563
ExViT-TS
0.8640
0.8492
0.8634
TSViT-SAR
0.8632
0.8472
0.8622
PSE-TAE-SAR
0.8624
0.8462
0.8610
Presto-SAR
0.8583
0.8413
0.8572
TABLE XI: Full Valid-Pixel Spatial Mapping Accuracy on the USA Dataset (Six-Band Input, 28,950,454 Valid Pixels)
Fig. 5: Comparison of full valid-pixel spatial crop classification maps on the USA dataset. All models use the six-band SAR time-series input and are rendered under the same CDL color legend.
Fig. 6: Comparison of full valid-pixel error distributions on the USA dataset. Gray denotes pixels matching the CDL reference labels, and red denotes misclassified pixels.
Task
Method
OA
Macro F1
Rice
From Scratch
0.9944
0.9883
With Pre-training
0.9963
0.9921
Wheat
From Scratch
0.9905
0.9902
With Pre-training
0.9923
0.9921
TABLE XII: Independent Binary Classification Results for Rice and Wheat on Chongming Island
Fig. 7: Monthly crop classification comparison on Chongming Island (11 classes). Left: Overall Accuracy (OA); right: Macro F1. The pre-trained model achieves higher OA in 9 out of 10 months, with pronounced Macro F1 gains during the March–April phenological transition period.
Category
F1 (From Scratch)
F1 (With Pre-training)
Δ F1
Test Samples
Rice
0.9967
0.9970
+0.03%
90,979
Wheat
0.9958
0.9962
+0.04%
52,824
Corn
0.9909
0.9914
+0.05%
3,570
Soybean
0.9812
0.9873
+0.61%
5,503
Vegetables
0.9840
0.9813
-0.27%
2,206
Rapeseed
0.9886
0.9886
0.00%
5,952
TABLE XIII: Per-Class F1 Scores on Chongming Island (Pooled Across 10 Months, 11 Classes, 225,970 Samples)
Fig. 8: Ablation comparison of 6-band fusion versus 3-band inputs. Left: Estonia dataset; right: USA dataset. Comparing 3-band (Backscatter), 3-band (Polarimetric), and 6-band (Fusion) confirms strong physical complementarity between the two feature groups.
Dataset
Input Configuration
OA
Macro F1
Weighted F1
Estonia
3-band (Backscatter)
0.9132
0.8854
0.9129
3-band (Polarimetric)
0.9296
0.9117
0.9273
6-band (Fusion)
0.9678
0.9572
0.9677
USA
3-band (Backscatter)
0.8222
0.7992
0.8204
3-band (Polarimetric)
0.8056
0.7817
0.8026
6-band (Fusion)
0.8729
0.8616
0.8723
TABLE XIV: Ablation Study on 6-Band vs. 3-Band Inputs (Pre-Trained Models). 3-Band (Backscatter): VV, VH, and VV/VH; 3-Band (Polarimetric): H/α/A .
Dataset
Variant
OA
Macro F1
Weighted F1
Estonia
A
0.9537
0.9363
0.9531
B
0.9615
0.9488
0.9612
C
0.9678
0.9572
0.9677
USA
A
0.8604
0.8471
0.8601
B
0.8676
0.8568
0.8669
C
0.8729
0.8616
0.8723
TABLE XV: Ablation Results of Dual-Branch Encoding and Cross-Attention Under Fixed Six-Band Input. A: Single-Concat; B: Dual-NoCross; C: Dual-Cross.
Fig. 9: Ablation comparison of masking strategies. Left: Chongming Island dataset; right: Estonia dataset. SpanPhysics achieves the highest OA and Macro F1 on both benchmarks.
Dataset
Masking Strategy
OA
Macro F1
Chongming
Random
0.9929
0.9825
Span (len=3)
0.9936
0.9805
SpanPhysics
0.9938
0.9894
Estonia
Random
0.9459
0.9298
Span (len=3)
0.9581
0.9465
SpanPhysics
0.9678
0.9572
TABLE XVI: Ablation Study on Masking Strategies (Masking Ratio pmask=0.15 , Span Length L=3 )
Fig. 10: Few-shot evaluation results on the Estonia dataset. Left: Overall Accuracy (OA); right: Macro F1. Pre-training consistently outperforms training from scratch across all label ratios, with amplified gains under the 1% and 5% label regimes.
Label Ratio
OA
Δ OA
Macro F1
Δ F1
From Scratch
With Pre-training
From Scratch
With Pre-training
1%
0.7736
0.7994
+2.58%
0.5710
0.6229
+5.19%
5%
0.8200
0.8395
+1.95%
0.7078
0.7397
+3.19%
10%
0.8584
0.8711
+1.27%
0.7877
0.8038
+1.61%
50%
0.9103
0.9215
+1.12%
0.8801
0.8956
+1.55%
100%
0.9561
0.9678
+1.17%
0.9413
0.9572
+1.59%
TABLE XVII: Few-Shot Evaluation Results on the Estonia Dataset
Fig. 11: Six-band Sentinel-1 temporal profiles for rice and wheat on Chongming Island. Solid curves denote class means, and shaded bands indicate interquartile ranges (IQR).
Fig. 12: Temporal evolution of the six SAR features within the exact same parcels transitioning from winter–spring wheat to summer–autumn rice.
Fig. 13: Temporal separability heatmap (absolute Cohen’s d ) across the six SAR bands between fixed wheat parcels and fixed rice parcels on Chongming Island. Higher values indicate stronger inter-class separability on a given acquisition date and feature.
Feature
Same-Month Rice vs. Wheat
Pooled Wheat vs. Rice Seasons
Exact Wheat–Rice Rotation Parcels
VV
0.073
0.039
0.205
VH
0.629
0.426
0.571
VV/VH
0.686
0.153
0.070
H
1.674
0.399
0.347
α
1.659
0.472
0.387
A
1.456
0.481
0.399
TABLE XVIII: Six-Band Separability (Absolute Cohen’s d ) Between Rice and Wheat Across Analytical Settings on Chongming Island