Transcript Intelligence

Conversation Detail

Technical discussion on dashcam hardware and ANPR

Dashcam evaluation and testing strategy

Behnam dashcam ยท 2026-06-03 12:04:54 to 2026-06-03 12:11:37

Review: Pending Review

Summary

The speaker, likely Behnam, discusses evaluating dashcams for ANPR (automatic number plate recognition). He compares the 70mai M310 with the Samsung S22+ and recommends testing by driving the same route with both, feeding footage through BITRA for vehicle detection, plate detection, and OCR. He mentions factors like bitrate (M310 at 7-11 Mbps), motion blur, and compression artifacts. He reviews alternatives: BlackVue (fleet-oriented, LTE, expensive), Viopo (excellent image quality, A229 Pro with 5GHz Wi-Fi), and Thinkware (reliable, less hacker-friendly). He specifically recommends the Viopo A329 for image quality. He advises against upgrading the Pi's Wi-Fi adapter until the M310's transfer issues (flaky hotspot, 503 errors, file locks) are addressed. Later, he speculates about using an Android device with the 70mai app to discover the exact commands used for file transfer.

Facts

  • M310 typical 60-second clip is 50-80 MB (7-11 Mbps).
  • BlackVue is designed for fleets, has LTE modules, cloud connectivity.
  • Viopo A229 Pro has 5 GHz Wi-Fi and SSD support.
  • Viopo A329 is recommended for image quality.
  • M310 Wi-Fi hotspot is flaky, web server returns 503, files locked during recording.

Tasks

  • Drive same route with M310 and S22+.
  • Feed footage through BITRA.
  • Count vehicle detections, plate detections, successful OCR reads.
  • Compare results between cameras.
  • Do not spend money on faster Wi-Fi adapter for Pi yet.

Issues

  • M310 Wi-Fi hotspot unreliable.
  • Camera web server returns 503 errors.
  • Camera locks files while recording.
  • Camera only supports 2.4 GHz Wi-Fi.
  • Transfer pipeline may be bottlenecked by camera, not Pi.