I followed the "Video Occupancy Sensor Source Build Guide for Outdoors Release" to try to build the VOS library from the sources so that I can trace it. I want o build a debug library. After finishing importing the vos_lib_project into the workspace, I can't compile the project. The "build" icon is simply grayed out. I can't Set Active in the Build Configurations for the same reason. I also tried to download the older version of CCES, but failed to locate them on the internet. I've also tried to run the pre build command vos.bat, but also failed. It seems to me that the path is not setup correctly to find the blkfn.exe. Maybe it is not supposed to be run manually, I don't know.
Any comment ?
I imported this code and build icon is not greyed out as you mentioned.
Regarding "I can't Set Active in the Build Configurations for the same reason". >> We would suggest you to refer the below CCES help path. CCES help: CrossCore® Embedded Studio 2.8.0 > Integrated Development Environment > Managing CCES Projects > Managing Build Configurations If you are still facing the same issue, please provide below details to assist you better on this issue. 1) Please let us know which CCES version and operating system that you are using. 2) Are you facing issue in specific project(vos_lib_project) or all the projects? 3) Can you please share us the screenshot of the issue you are facing?
Thank you again for answering my questions.
When I tried it again and followed the instructions in KT-1645_VOS_Outdoors_SourceBuildGuide.pdf and tried to build a debug library, and it worked this time. yet, I still can't trace into
nRetVal = gpVosHndl->pProcess(gpVosHndl, &goInputData, (void*)ADI_VOS_NULL, &goOutputData, (void*)&goVosOutput);
This is in the function of adi_RunVOS(). or other similar library calls. In all cases, it shows me error messages like :
Can't find a source file at "C:\Jenkins\workspace\CCES2_8_0_build_compiler_and_libs_bf\DSPLIB\LIBC\src\platform\frio\crt\sec_int_dispatcher.asm"
I couldn't locate this file sec_int_dispatcher.asm anywhere in the project tree.
I am assuming BLIP2 uses some type of Convolutional Neural Network for the car and human detection. My eventual goal is to try to understand your sensing algorithm and see if I can retrain the CNNs to detect some other objects, such as cups, cell phone, books, etc. Is this possible ?
Thank you very much again.
Regarding "I couldn't locate this file sec_int_dispatcher.asm anywhere in the project tree." >> Please follow the below steps to locate the file: 1) Click "Locate File..." as in the screenshot.
2) Enter "[Installation directory]\Analog Devices\CrossCore Embedded Studio 2.8.0\Blackfin\lib\src\libc\crt" in the "File Name" textbox. If you have installed the CCES in the different location,please change the path accordingly.
3) Click Open and choose "sec_int_dispatcher.asm".
4) Click Open Regarding "I am assuming BLIP2 uses some type of Convolutional Neural Network for the car and human detection. My eventual goal is to try to understand your sensing algorithm and see if I can retrain the CNNs to detect some other objects, such as cups, cell phone, books, etc. Is this possible ?" >> I am not an expert in this algorithm, meanwhile I would suggest you to refer the release note which is available in the below mentioned path. [Installation directory]\Analog Devices\SoftwareModules\VideoOccupancySensor-BF-Rel3.3.0\ReleaseNotes.pdf
Thank you very much for your time on showing me the location of the file. When I searched for that particular file, I used git bash to do "grep -nirl sec_int_dispatcher.asm" under the folder of C:\Analog Devices\SoftwareModules, without considering that the file might be located in IDE folder. I am sorry for my ignorance.
I know you said that you are not familiar with this VOS algorithm, but since you work in this company, can you ask the people working in the VOS team to provide some information on the question I asked ? I have read the release notes and there is nothing there to help me get the answer. My question is very simple that can I train the existing VOS algorithm to detect objects other than human and vehicle ?
This is a yes or no question.
Thank you for your time.