search

DoubangoTelecom / ultimateALPR-SDK

World's fastest ANPR / ALPR implementation for CPUs, GPUs, VPUs and NPUs using deep learning (Tensorflow, Tensorflow lite, TensorRT, OpenVX, OpenVINO). Multi-Charset (Latin, Korean, Chinese) & Multi-OS (Jetson, Android, Raspberry Pi, Linux, Windows) & Multi-Arch (ARM, x86).
https://www.doubango.org/webapps/alpr/
Other
617 stars 158 forks source link

disable compv info #148

Closed scizors closed 3 years ago

scizors commented 3 years ago

Hi, At the moment I am testing the trial version ultimateALPR-SDK in my ARM machine with German number plates C++ solution. I am using snapshot of the car numberplates inside Germany As I am getting with most of the number plates output as

*[ULTALPR_SDK INFO]: result: {"duration":4353,"frame_id":0}

I would like to know few things

  1. Differnce between the trial version and commerical paid version in terms of accuracy (apart from last character in the number plate is *)
  2. As I see there is [COMPV INFO]: and [ULTALPR_SDK INFO]: is it possible to disable *[COMPV INFO]: in my trial version just for testing purpose.
DoubangoTelecom commented 3 years ago

1/ same accuracy. You should attach sample plates the SDK fails to process. 2/ https://www.doubango.org/SDKs/anpr/docs/Configuration_options.html#debug-level

scizors commented 3 years ago

Hi,

Thank you for the info please find it. I have lots of image file with same situation which the algorithm fails

  1. I have changed to "warn" as shown below

[UltAlprSdkEngine] jsonConfig: {"debug_level": "warn","debug_write_input_image_enabled": false,"debug_internal_data_path": ".","num_threads": -1,"gpgpu_enabled": true,"klass_vcr_gamma": 1.5,"detect_roi": [0, 0, 0, 0],"detect_minscore": 0.1,"car_noplate_detect_min_score": 0.8,"pyramidal_search_enabled": true,"pyramidal_search_sensitivity": 1.0,"pyramidal_search_minscore": 0.3,"pyramidal_search_min_image_size_inpixels": 800,"recogn_minscore": 0.3,"recogn_score_type": "min","assets_folder": "../../../assets","charset": "latin","recogn_rectify_enabled": false,"car_noplate_detect_enabled": false,"ienv_enabled": false,"openvino_enabled": true,"openvino_device": "CPU","klass_lpci_enabled": false,"klass_vcr_enabled": false,"klass_vmmr_enabled": false,"klass_vbsr_enabled": false}

Still I am seeing *[COMPV INFO]: is there any waw to diable?

DoubangoTelecom commented 3 years ago
DoubangoTelecom commented 3 years ago

I have checked with https://www.doubango.org/webapps/alpr/ and it fails to recognize the plate. These kind of images should be easily recognized. I'm reopening this ticket and we'll check. If you can share with us more images to use in our training set then it'll help us improve the algo.