/j8ddlmZddlmZddlmZddlZddlZddl m Z ddl m Z m Z ddlmZmZddlmZdd lmZGd d e ZdS) ) annotations)Path)AnyN)DetectionValidator)LOGGERops) OBBMetrics batch_probiou)TorchNMS) plot_imagescveZdZdZd&d'fd Zd(fd Zd)dZd*fd Zd+dZd,dZ d-dZ d.d"Z d/d#Z d0d%Z xZS)1 OBBValidatoraA class extending the DetectionValidator class for validation based on an Oriented Bounding Box (OBB) model. This validator specializes in evaluating models that predict rotated bounding boxes, commonly used for aerial and satellite imagery where objects can appear at various orientations. Attributes: args (dict): Configuration arguments for the validator. metrics (OBBMetrics): Metrics object for evaluating OBB model performance. is_dota (bool): Flag indicating whether the validation dataset is in DOTA format. Methods: init_metrics: Initialize evaluation metrics for YOLO. _process_batch: Process batch of detections and ground truth boxes to compute IoU matrix. _prepare_batch: Prepare batch data for OBB validation. _prepare_pred: Prepare predictions for evaluation against ground truth. plot_predictions: Plot predicted bounding boxes on input images. pred_to_json: Serialize YOLO predictions to COCO json format. save_one_txt: Save YOLO detections to a txt file in normalized coordinates. eval_json: Evaluate YOLO output in JSON format and return performance statistics. Examples: >>> from ultralytics.models.yolo.obb import OBBValidator >>> args = dict(model="yolo26n-obb.pt", data="dota8.yaml") >>> validator = OBBValidator(args=args) >>> validator(model=args["model"]) N _callbacks dict | NonereturnNonect||||d|j_t |_dS)aInitialize OBBValidator and set task to 'obb', metrics to OBBMetrics. This constructor initializes an OBBValidator instance for validating Oriented Bounding Box (OBB) models. It extends the DetectionValidator class and configures it specifically for the OBB task. Args: dataloader (torch.utils.data.DataLoader, optional): DataLoader to be used for validation. save_dir (str | Path, optional): Directory to save results. args (dict, optional): Arguments containing validation parameters. _callbacks (dict, optional): Dictionary of callback functions to be called during validation. obbN)super__init__argstaskr metrics)self dataloadersave_dirrr __class__s d/home/longshao/multi-rider-rag/.venv/lib/python3.11/site-packages/ultralytics/models/yolo/obb/val.pyrzOBBValidator.__init__.s; XtZ@@@ !|| modeltorch.nn.Modulect||j|jjd}t |tod|v|_d|j _ dS)zInitialize evaluation metrics for YOLO obb validation. Args: model (torch.nn.Module): Model to validate. DOTArN) r init_metricsdatagetrsplit isinstancestris_dotaconfusion_matrixr)rr valrs rr%zOBBValidator.init_metrics>sa U###immDIOR00!#s++=#  %*"""rpredsdict[str, torch.Tensor]batchdict[str, np.ndarray]c|djddks|djddkr5dtj|djd|jftiSt |d|d}d||d|d|iS)aCompute the correct prediction matrix for a batch of detections and ground truth bounding boxes. Args: preds (dict[str, torch.Tensor]): Prediction dictionary containing 'cls' and 'bboxes' keys with detected class labels and bounding boxes. batch (dict[str, torch.Tensor]): Batch dictionary containing 'cls' and 'bboxes' keys with ground truth class labels and bounding boxes. Returns: (dict[str, np.ndarray]): Dictionary containing 'tp' key with the correct prediction matrix as a numpy array with shape (N, 10), which includes 10 IoU levels for each detection, indicating the accuracy of predictions compared to the ground truth. Examples: >>> preds = {"cls": torch.randint(0, 5, (100,)), "bboxes": torch.rand(100, 5)} >>> batch = {"cls": torch.randint(0, 5, (50,)), "bboxes": torch.rand(50, 5)} >>> correct_matrix = validator._process_batch(preds, batch) clsrtpdtypebboxes) shapenpzerosniouboolr match_predictionscpunumpy)rr.r0ious r_process_batchzOBBValidator._process_batchIs& < a A % %u);A)>!)C)C"(E%L$6q$949#ETRRRS SE(OU8_==d,,U5\5<MMQQSSYY[[\\r torch.Tensorlist[dict[str, torch.Tensor]]ct|}|D]6}tj|d|dgd|d<7|S)zPostprocess OBB predictions. Args: preds (torch.Tensor): Raw predictions from the model. Returns: (list[dict[str, torch.Tensor]]): Processed predictions with angle information concatenated to bboxes. r7extradim)r postprocesstorchcatpop)rr.predrs rrIzOBBValidator.postprocessasb##E** T TD"YX8I8I'JPRSSSDNN rsiintdict[str, Any]c|d|k}|d|d}|d|}|d|}|djdd}|d |}|jd r@|d dd ftj||j gd||||||d|dS)a~Prepare batch data for OBB validation with proper scaling and formatting. Args: si (int): Sample index within the batch. batch (dict[str, Any]): Dictionary containing batch data with keys: - batch_idx: Tensor of batch indices - cls: Tensor of class labels - bboxes: Tensor of bounding boxes - ori_shape: Original image shapes - img: Batch of images - ratio_pad: Ratio and padding information Returns: (dict[str, Any]): Prepared batch data with scaled bounding boxes and metadata. batch_idxr3rFr7 ori_shapeimgN ratio_padr.)device)rrYrim_file)r3r7rSimgszrVrZ)squeezer8mul_rJtensorrX) rrNr0idxr3bboxrSr[rVs r_prepare_batchzOBBValidator._prepare_batchos K B&El3''++Xs#+&r* e "122&+&r* 9Q< V bqbM  u|E$+FFF|||T U U U""Y'+    rnic 8sdStD]%\}}tj|d|z|d<&d}fd|D}t |d||d|jd|d z |j|j dS) aPlot predicted bounding boxes on input images and save the result. Args: batch (dict[str, Any]): Batch data containing images, file paths, and other metadata. preds (list[dict[str, torch.Tensor]]): List of prediction dictionaries for each image in the batch. ni (int): Batch index used for naming the output file. Examples: >>> validator = OBBValidator() >>> batch = {"img": images, "im_file": paths} >>> preds = [{"bboxes": torch.rand(10, 5), "cls": torch.zeros(10), "conf": torch.rand(10)}] >>> validator.plot_predictions(batch, preds, 0) NconfrRrcXi|]%tjfdDd&S)c g|] }| Srg).0xks r z..s&;&;&;qt&;&;&;rrrG)rJrK)rhrjr.s @r z1OBBValidator.plot_predictions..s@RRREI&;&;&;&;U&;&;&;CCCRRRrrTrZ val_batchz _pred.jpg)imageslabelspathsfnamenameson_plot) enumeraterJ ones_likekeysr rrrrs)rr0r.rbirMrv batched_predss ` rplot_predictionszOBBValidator.plot_predictionss  F '' B BGAt %V = = AD  Qx}}RRRRTRRR <  "-";b";";";;*L       rprednpbatchc t|d}|j}|rt|n|}|d}t j|dd}t|||d|dD]g\}} } } |j ||j |j t| t| dd|Dd | Dd hd S) aConvert YOLO predictions to COCO JSON format with rotated bounding box information. Args: predn (dict[str, torch.Tensor]): Prediction dictionary containing 'bboxes', 'conf', and 'cls' keys with bounding box coordinates, confidence scores, and class predictions. pbatch (dict[str, Any]): Batch dictionary containing 'imgsz', 'ori_shape', 'ratio_pad', and 'im_file'. Notes: This method processes rotated bounding box predictions and converts them to both rbox format (x, y, w, h, angle) and polygon format (x1, y1, x2, y2, x3, y3, x4, y4) before adding them to the JSON dictionary. rZr7rFrdr3c.g|]}t|dSroundrhris rrkz-OBBValidator.pred_to_json.. 444QU1a[[444rc.g|]}t|dSrrrs rrkz-OBBValidator.pred_to_json..rr)image_id file_name category_idscorerboxpolyN)rstem isnumericrOrxywhr2xyxyxyxyviewziptolistjdictappendname class_mapr) rrzr{pathrrrrrbscs r pred_to_jsonzOBBValidator.pred_to_jsons4F9%&&y $ 0 0:3t999dX!$'',,R33dkkmmT[[]]E&M#a& "1a[[44!44444!444       r save_confr<r8tuple[int, int]filerc Pddl}ddlm}||j|d|df|jd|jt j|d|dd|d dgd   || dS) aSave YOLO OBB detections to a text file in normalized coordinates. Args: predn (dict[str, torch.Tensor]): Prediction dictionary containing 'bboxes', 'conf', and 'cls' keys with bounding box coordinates (including angle), confidence scores, and class predictions. save_conf (bool): Whether to save confidence scores in the text file. shape (tuple[int, int]): Original image shape in format (height, width). file (Path): Output file path to save detections. Examples: >>> validator = OBBValidator() >>> predn = { ... "bboxes": torch.tensor([[100, 100, 50, 30, 45]]), ... "conf": torch.tensor([0.9]), ... "cls": torch.tensor([0]), ... } >>> validator.save_one_txt(predn, True, (640, 480), Path("detection.txt")) rN)ResultsrYr5r7rdrFr3rG)rrrr)r) r?ultralytics.engine.resultsrr:uint8rrrJrK unsqueezesave_txt)rrzrr8rr9rs r save_one_txtzOBBValidator.save_one_txts& 666666 BHeAha) : : :* 5?E&M,C,CB,G,GuI_I_`bIcIcdjklll    (49( - - - - -rc i|dtj|d|d|d|ddiS)z.Scales predictions to the original image size.r7r[rSrVT)rVxywh)r scale_boxesclone)rrzr{s r scale_predszOBBValidator.scale_predss`  cowx!6!6!8!8&:MY_`kYlsw   rstatsc |jjr|jrt|jrddl}ddl}ddlm}|j dz }|j dz }| dd| t|}tjd|d |D]}|d } |d } |j|d d z dd} |d} t|d| z ddd5} | | d| d| dd| d d| dd| dd| dd| dd| dd| dddddn #1swxYwY|j dz }| dd|t$}tjd|d |D]}|d dd d} |d }d!|||d dd"D\}}|d#|d |d d z }} }|dxx|z cc<|d xx|z cc<|| |g|| ||D](\} }t3j|}t3j|ddddfdz}|ddddf|z}|dddf}|ddddf}|ddddfxx|z cc<t=j||d$t@%}||}tCj"|ddddf#d&d'}t3j$||ddddfgd&(%D]}|jtM|d&dd} d)|dd*D} tO|d*d} t|d| z ddd5} | | d| d| dd| d d| dd| dd| dd| dd| dd| dddddn #1swxYwY*|S)+zEvaluate YOLO output in JSON format and save predictions in DOTA format. Args: stats (dict[str, Any]): Performance statistics dictionary. Returns: (dict[str, Any]): Updated performance statistics. rN) defaultdictzpredictions.jsonpredictions_txtT)parentsexist_okz'Saving predictions with DOTA format to z...rrrrY -rTask1_z.txtazutf-8)encodingrUrrWr~ predictions_merged_txtz.Saving merged predictions with DOTA format to __z \d+___\d+c34K|]}t|VdS)N)rO)rhrs r z)OBBValidator.eval_json..s([[1A[[[[[[r___rg333333?)iou_funcrFr}rGc.g|]}t|dSrr)rhrws rrkz*OBBValidator.eval_json..+s 555q!555r)(r save_jsonr+lenrjsonre collectionsrrmkdirloadopenrinforrreplace writelineslistr(compilefindallextendritemsrJr^maxitemrr fast_nmsr rrrrKrrOr)rrrrr pred_jsonpred_txtr&drr classnamepfpred_merged_txtmerged_resultspatternriyr`r3max_whrscoresrrws r eval_jsonzOBBValidator.eval_jsonsP 9 4 v4<4 vC OO4 v KKK III / / / / / / (::I}'88H NN4$N 7 7 799T)__--D KO(OOO P P P r rZ='  Jq'7!';<DDS#NN fIX(<(<(<<BBBCRYZZZr^_LLH!p!pu!p!pqt!p!pad!p!pQqT!p!pAaD!p!pSTUVSW!p!pZ[\]Z^!p!pabcdae!p!phijkhl!p!p!pqqqrrrrrrrrrrrrrrr#m.FFO  ! !$ ! > > >([..N K]]]] ^ ^ ^ 6 6Z=..tQ77:**\22[[ 7AjM(J(J1(M(S(STY(Z(Z[[[1#$V9aj!M:JQ:NSeQ1 Q1  UCL)))x(//5555"0"6"6"8"8 v v$|D))42A2;//4466:AaCL6)aaadBQBK%%''!!!RaR%A %a}MMMAw&tAAArrE{3388Q??AtAAAqsF|#4"===DDFFvvA $ 3qu:: 6 > >sC H HI55af555A!!B%OOE3GI3G3G!GMMMs]deeevij %t%t5%t%t1Q4%t%t!A$%t%t1%t%tPQRSPT%t%tWXYZW[%t%t^_`a^b%t%tefghei%t%tlmnolp%t%t%tuuuvvvvvvvvvvvvvvv v s&?A$E//E3 6E3 /A$SS# &S# )NNNN)rrrr)r r!rr)r.r/r0r/rr1)r.rBrrC)rNrOr0rPrrP)r0rPr.rCrbrOrr)rzr/r{rPrr) rzr/rr<r8rrrrr)rzr/r{rPrr/)rrPrrP)__name__ __module__ __qualname____doc__rr%rArIraryrrrr __classcell__)rs@rrrs6$$$$$$$ + + + + + +]]]]0          B    :<....<    ????????rr) __future__rpathlibrtypingrr?r9rJultralytics.models.yolo.detectrultralytics.utilsrrultralytics.utils.metricsr r ultralytics.utils.nmsr ultralytics.utils.plottingr rrgrrrs#""""" ======))))))))????????******222222_____%_____r