Photo capture stops mid-flight?

Recently we attempted to capture some imagery with our Sequoia camera using the Atlas app. The Sequoia is attached to our DJI Phantom 3 Professional with the MicaSense integration kit designed for that purpose. However, we’ve had a number of issues. This is issue 2 of 3.

I’m rather desperate to get this camera operational as the area we are trying to photograph is about to be destroyed by construction work, and after 1.5 weeks without response from Parrot technical support after creating a support case, I’m hoping that some of you may have sufficient experience with Sequoia to offer a solution or some advice.

After configuring our flight plans and camera options in the Atlas app, the mission is uploaded to our drone and the flight is flown. However, during the flight only a few (a couple dozen) photographs get successfully captured (the flight should have produced something like 300 photos [1200 TIFF files]). This happens whether we are using an SD card, or writing to internal memory.

The sequoa_diag.dat file shows a number of times that SICILIA_ERROR messages are thrown, but it appears that the system continues to work, and indeed a number of photographs are captured and stored. However, after a number (the exact number varies widely) of photos, the following errors appear in the log and no further photographs are captured:

HAL_i2c_msg_send() : 121, Remote I/O error
one or more i2c addresses are not correct
libHAL’s HAL_read() : error send
HAL_thread_handler: ERROR reading sensors (Error 2)
HAL_i2c_msg_send() : 121, Remote I/O error
one or more i2c addresses are not correct
libHAL’s HAL_read() : error send
HAL_i2c_msg_send() : 121, Remote I/O error
one or more i2c addresses are not correct
HAL_thread_handler: ERROR reading sensors (Error 2)
Gps disconnected, libgpsman not launched, aborting
GpsReader no uart connection
GpsReader thread killed
Powering off gps…
HAL_i2c_msg_send() : 19, No such device
libHAL’s HAL_read() : error send

This starts a series of issues in the log and eventually a crash of sensors_IR:
Memory section “colibry_sensors_IR” successfully open with revision number : 4
fd 7 not found in loop 0x76a09370
EventsFsm: void Libfsm::EventsFsm::stop() [with EventType = Sicilia::controller::eControlEvents] stop
EventsFsm: destructor
EventsFsm: void Libfsm::EventsFsm::stop() [with EventType = Sicilia::eEstimatorEvents] stop
worker 417 started
unw_get_proc_name failed: -1
unw_get_proc_info failed: -10
no region containing address 0 found
*** *** *** *** *** *** CRASH *** *** *** *** *** *** *** ***
exe : /usr/bin/sensors_ir

Has anyone ever had their Sequioa just stop taking pictures mid-flight? Have you ever seen such I/O errors in your log files? Do you have an suggestions on what might be the cause? Or what could be a solution?

The entire diagnostics log is available for review at:

My Sequoia stopped capturing images a few times from overheating. However, it currently doesn’t capture any images at all any more (except radiometric calibration images), and I’m stumped… How did you access the log?

I seem to have the same SICILIA error. I managed to access the log. Any idea how to solve it? It all seems to be an issue of firmware 1.7.1, but the Sequoia doesn’t let me downgrade…