Commit Graph
89 Commits
Author SHA1 Message Date
nghianguyen 91d81cbd88 fix onnx yolo race condition 2026-05-04 23:45:05 +10:00
nghianguyen 03a63142ec Disable stabilized feature by default. Need ,more time to test 2026-05-04 09:39:07 +10:00
nghianguyen cff2a90fad Support multiple GPUs 2026-05-02 12:24:53 +10:00
nghianguyen fc9cf9671b Fix ANSONNXYOLO:
1. core/ANSLicensingSystem/ANSLicense.h:268 — added default initializers to ModelConfig POD members (inpWidth = 0, inpHeight = 0, etc.). This was the actual bug: CreateANSODHandle declares a local ModelConfig and only sets a few fields, leaving inpWidth/inpHeight as indeterminate stack memory. On your machine that memory consistently held ~80, slipping past the <=0 → 640 guard and driving the ORT session at 80×80 — which crashes YOLOv11's /model.12/Concat because 2·ceil(80/32)=6 ≠ ceil(80/16)=5.

2. modules/ANSODEngine/ANSONNXYOLO.h:75 — ONNXYOLO::setInputShape() now snaps width/height up to the next multiple of 32 (with a 64-pixel floor). Defense in depth: even if a future caller hands in a non-stride-32 size, the FPN concat alignment is preserved.

3. modules/ANSODEngine/ANSONNXYOLO.cpp — two diagnostic additions:

WarmUpEngine bails with a clear log line if the warm-up dims aren't a multiple of 32, instead of letting the model.12 error fire and mask the real cause.
transform() logs the input shape, canvas shape, source size, and dynamic flag once per session in DebugView — so any future shape-mismatch crash points at the offending HxW directly.
2026-05-01 20:24:08 +10:00
nghianguyen 1c8abf2a88 Fix AWS MinIO 2026-05-01 11:16:47 +10:00
nghianguyen 91efbb9a8f Fix AMD ALPR_OCR and custom model 2026-05-01 07:53:18 +10:00
nghianguyen e1f37ec767 Add ANSALPR country format 2026-04-28 19:23:46 +10:00
nghianguyen cd5d6f1923 Remove spaces, ".", and "-" from ALPR 2026-04-28 17:49:37 +10:00
nghianguyen 234f2c68a2 bucket prewarm + batch padding 2026-04-28 17:12:27 +10:00
nghianguyen dcf974c35c Fix AMD by changing from GetTensorData<T>() to GetTensorMutableData<T>() 2026-04-28 13:25:02 +10:00
nghianguyen da94e7f6da strip sensitive information from logs 2026-04-26 07:35:30 +10:00
nghianguyen bab94bf9bc Add log and change logger policy 2026-04-25 20:29:58 +10:00
nghianguyen ef2a122fec Add logs to custom models and model optimisation 2026-04-25 13:17:36 +10:00
nghianguyen 8a95ed6b8c add debug 2026-04-25 09:44:55 +10:00
nghianguyen 3a527d2dad Remove locks 2026-04-24 17:10:29 +10:00
nghianguyen fd2394a85a Fix deadlock 2026-04-24 12:54:16 +10:00
nghianguyen e2bf17289d Fix model extract race issue to all classes 2026-04-24 12:19:54 +10:00
nghianguyen baa88bcc48 Fix Concurrent extract + open race on train_last.onnx 2026-04-24 11:29:28 +10:00
nghianguyen 91bdb3f96b Fix play once video player 2026-04-23 08:36:52 +10:00
nghianguyen c625898f61 Fix Video Player that support generate black image when finish playing video (reach the end of video) 2026-04-23 08:08:34 +10:00
nghianguyen 6135649835 Use yuv and fix FR to use batch 4 2026-04-22 17:11:17 +10:00
nghianguyen a0d5c81814 Update 2026-04-22 16:52:40 +10:00
nghianguyen fc47a42c6a Use yuv and fix FR to use batch 4 2026-04-22 16:01:44 +10:00
nghianguyen 24d6d2e0a5 Use tracker and voting system to correct ALPR 2026-04-22 11:19:42 +10:00
nghianguyen 0e203eef99 Fix double stops 2026-04-22 10:19:03 +10:00
nghianguyen 57cc8e0a56 Fix double stop in ANSVideoPlayer 2026-04-22 10:10:16 +10:00
nghianguyen 97d814936d Enable log information. Disable NPU in U9 2026-04-21 15:48:27 +10:00
nghianguyen 7e772f76bc Improve ANSCV 2026-04-21 09:26:02 +10:00
nghianguyen 9f0a10a4c8 Improve ANSCV with sotfware decoder:
Thread-local staging Mat (video_player.cpp:1400-1407) — single biggest win. Eliminates the 12 MB per-call malloc/free cycle.
Contiguous get_buffer2 allocator (video_decoder.cpp:35-102) — keeps the 3 bulk memcpys cache-friendly. Would also enable FAST/zero-copy for resolutions where visible_h % 64 == 0.
SW-decoder thread config (video_decoder.cpp:528-540) — thread_count=0, thread_type=FRAME|SLICE. FRAME is downgraded to SLICE-only by AV_CODEC_FLAG_LOW_DELAY, but decode throughput is sufficient for your input rate.
SetTargetFPS(100) delivery throttle (already there) — caps onVideoFrame post-decode work at 10 FPS. Keeps the caller path warm-cached.
Instrumentation — [MEDIA_DecInit] / [MEDIA_Convert] / [MEDIA_SWDec] / [MEDIA_Timing] / [MEDIA_JpegTiming] — always-on regression detector, zero cost when ANSCORE_DEBUGVIEW=OFF.
2026-04-20 12:18:43 +10:00
nghianguyenandClaude Opus 4.7 adf32da2a2 Add early stale-out in ANSRTSPClient::GetImage
When the decoder hasn't produced a frame in 5s, skip the call to
_playerClient->getImage() entirely and return the cached frame with
unchanged _pts. LabVIEW sees STALE PTS one poll earlier and can
trigger reconnect sooner.

Threshold matches the existing checks on the duplicate-PTS branch and
in areImagesIdentical() so all three stale paths agree. Near-zero cost:
one getLastFrameAgeMs() call before the main path.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-20 08:35:23 +10:00
nghianguyen 3418090042 - Remove ALPR_OD Tracker and voting system.
- Disable debugview
2026-04-19 23:15:50 +10:00
nghianguyen 51fe507dfd Fix unregister issue 2026-04-19 14:47:29 +10:00
nghianguyen 2de11c4b0c Fix: remove the leak protection entirely 2026-04-18 20:51:50 +10:00
nghianguyen 247c91db50 Fix ANSAWS3 multiple instance creation 2026-04-18 20:33:44 +10:00
nghianguyen a47ea25e8b Fix Image management (BMP, BMPtoJPEG) and remove debug info from AWS Upload 2026-04-17 20:59:21 +10:00
nghianguyenandClaude Opus 4.7 fcc985efb9 Fix AWS S3 upload failure: remove ClearAllHeaders that was wiping SigV4 state
Removing the Fix 1 ClearAllHeaders/ClearAllQueryParams block from all 8
upload entry points. CkRest.AddHeader already overwrites by header name,
so the clears were unnecessary; they also wipe the internal state
Chilkat's SigV4 signer relies on between requests, which caused AWS to
reject every PUT with HTTP 400 MaxMessageLengthExceeded (max 35240 B).
Per-step ClearAllQueryParams inside multipart kept (followed immediately
by AddQueryParam).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-17 08:13:32 +10:00
nghianguyen a63feab0ff Ensure the threadsafe for video upload 2026-04-17 07:03:03 +10:00
nghianguyen 2abfd6e87c Update the video conversion: Ensure the correct file order 2026-04-16 22:32:03 +10:00
nghianguyen 7cb0b3930e Disabled BMP for now 2026-04-16 16:34:27 +10:00
nghianguyen 8bb4f49b09 Improve jpeg conversion 2026-04-16 10:59:28 +10:00
nghianguyenandClaude Opus 4.6 6c72751a14 Add NvJpegPool (4 encoders) and JPEG passthrough in BmpToJpeg
- NvJpegPool: singleton pool of 4 NvJpegCompressor instances with
  lock-free slot acquisition (~160MB VRAM). Threads that can't grab
  a slot fall back to TurboJPEG with zero wait.
- JPEG passthrough: BmpToJpeg now checks if input is already JPEG
  (FF D8 FF magic) and copies directly without re-encoding.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-16 08:33:17 +10:00
nghianguyenandClaude Opus 4.6 53a82da74a Add nvJPEG GPU-accelerated JPEG encoding with NVIDIA auto-detection
BmpToJpeg was slow (~25-45ms for 4K) due to two bottlenecks:
1. cv::imdecode for BMP parsing (unnecessary for uncompressed BMP)
2. TurboJPEG CPU encoding (~11ms for 4K)

Fix 1: Zero-copy BMP parsing — parse header directly and wrap pixel
data in cv::Mat without allocation or copy. Eliminates ~47MB of heap
allocations per 4K frame.

Fix 2: NvJpegCompressor class using nvJPEG hardware encoder on NVIDIA
GPUs (~1-2ms for 4K). Integrated into CompressJpegToString so all 5
JPEG encoding callsites benefit automatically. Reusable GPU buffer
avoids per-frame cudaMalloc/cudaFree. Silent fallback to TurboJPEG
on Intel/AMD or if nvJPEG fails.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-16 07:50:13 +10:00
nghianguyen 808df4656d Fix ALPR Batch and memory leak 2026-04-15 23:00:19 +10:00
nghianguyen b05c49ad93 Fix ALPR Batch and memory leak 2026-04-15 09:23:05 +10:00
nghianguyen 7778f8c214 Two-part fix
Fix 1 — Chunk oversized bucket groups (the correctness fix)
ONNXOCRRecognizer::RecognizeBatch now slices each bucket group into chunks of ≤ kRecMaxBatch before submitting to TRT. A frame with 30 crops in bucket 320 produces two back-to-back batched calls (24 + 6), both within the profile, both on the fast path.

Fix 2 — Raise the profile max from 16 to 24 (the performance fix)
The old profile max was 16; your real scenes routinely hit 24. Raising the profile max to 24 means the common 12-plate scene (24 crops) fits in a single batched call with no chunking needed. Scenes with > 24 crops now use chunking, but that's rare.
2026-04-15 07:27:55 +10:00
nghianguyen 5706615ed5 Support tracker to improve ALPR_OCR 2026-04-14 21:18:10 +10:00
nghianguyen f9a0af8949 Improve ALPR_OCR peformance 2026-04-14 20:30:21 +10:00
nghianguyen 3349b45ade Update video conversion 2026-04-13 21:55:57 +10:00
nghianguyen 8e60126c4c Fix mutex lock issues (OCR and FR) 2026-04-13 20:38:40 +10:00
nghianguyen 844d7396b2 Fix mutex lock issues 2026-04-13 19:48:32 +10:00