標籤: Intel Real Sense

RealSense 學習筆記/教學/分享（六）：PyQT製作介面GUI

發表於 由Jay Hung撰寫

GUI: graphic user interface

要給其他同事使用務必要做成有介面 所以就在主要程式都完成後，開始這部分

我做了兩版，Tkinter跟PyQt

Tkinter真的太2000年了 雖然是內建很方便，但是就是要用新一點的

最後做出來的是如下：

GitHub https://github.com/soarwing52/RealsensePython/blob/master/proccessor.py

繼續閱讀 “RealSense 學習筆記/教學/分享（六）：PyQT製作介面GUI" →

專案簡介：深度相機的應用 從資料蒐集到資料應用 萬能的Python呀 請賜與我神奇的力量！

發表於 由Jay Hung撰寫

這是我這半年來主要的大專案，中間用了許多不同的組成，分成許多小專案開發

目前公司的主要業務為道路調查，類似街景車的在各道路、產業道路拍照

而本案的需求是公司想要有可以測量的照片

從硬體選購-深度相機程式開發與測試-實地測試-視覺化GUI製作由我獨立完成

中間有使用到的一些程式庫有：

電腦視覺 opencv
GPS Port應用
Realsense 程式庫
PyQt, Tkinter的GUI
Flask app
ArcGIS/QGIS 插件製作

繼續閱讀 “專案簡介：深度相機的應用 從資料蒐集到資料應用 萬能的Python呀 請賜與我神奇的力量！" →

Roadmap: Real Sense Application

發表於 由Jay Hung撰寫

The RealSense application is a solo project combined many different components I created during my job in the road surveying company

As the idea of the project is to let colleagues can measure objects such as road width, sidewalk, and even cracks.

Since the implementation can’t simply program through it, this project has:

Hardware selection, Computer Vision, Location Service, Depth Camera Library, GUI visualization

Further development with

ArcGIS/QGIS plugin creation, Flask web-framework

繼續閱讀 “Roadmap: Real Sense Application" →

The RealSense application is a solo project combined many different components I created during my job in the road surveying company

As the idea of the project is to let colleagues can measure objects such as road width, sidewalk, and even cracks.

Since the implementation can’t simply program through it, this project has:

Hardware selection, Computer Vision, Location Service, Depth Camera Library, GUI visualization

Further development with

ArcGIS/QGIS plugin creation, Flask web-framework

繼續閱讀 “Roadmap: Real Sense Application" →

Raspberry Pi + Realsense: Flask server

發表於 由Jay Hung撰寫

I would start first with the easy part, Flask

Flask is a micro framework to set up at server, app

and without setup like the whole Django file structure framework.

The good part is the variable they use is quite similar, because template of jinja2 and django is very similar

the code: https://github.com/soarwing52/Remote-Realsense/blob/master/flask_server.py 繼續閱讀 “Raspberry Pi + Realsense: Flask server" →

Project introduction: Tablet control Raspberry Pi with Realsense Depth Camera

發表於 由Jay Hung撰寫

The whole Realsense D435 project started long ago, I was working with it the whole year actually, and I have more records

I started on my working laptop, then need to run it on old road survey laptop, which is a lot weaker
and with a long usb cable connecting from the laptop on the side seat to the camera set on the front top of the car, it is just not as satisfying.
As the new pi has usb3 port, this is a good choice to update the things a bit.
the final result is on Github

Hardware Components

Raspberry pi 4 4GB
Realsense D435
GPS reciever (BU353S4 in my case)
(I’m still working on how to set it up together properly)
Since the goal is to use a tablet control pi remotely without cables and a laptop sitting on the side seat, some key are required
1: the connection: after asking and googling, I started with tcp sockets, and end up with Flask REST API
2:Android App: at first I though I might have to program in Java for this app for the sockets, luckily I got away with it with Flask and the MIT App Inventor this is a website to create apps with block code languages.
3.Since it can be used as HTML, how to stream the video? MJPEG is the answer I found, for streaming IP camera was the keyword
The rest I can easily work with my scripts and my python skills
————————————————————————————————————————

The Interface

this is made with the website MIT App inventor, the start will initialize the whole threads
restart is to separate every street while surveying, because one street can be on file
Auto mode will take every 15 meters one shot, instead of manually doing it
Drop list is to set the distance, not neccesarry to be 15 meters
the map opens in another page, but currently the GPS locator and how to show more informations are too complicated, still working
————————————————————————————————————————-

The Structure

The front end will easily send GET to the Flask app, and I just define functions, just as simple as any GUI
such as:
@app.route('/auto/')
def auto(in_text):
    a.command = in_text
    return in_text

in this part I used dynamic link, so I don’t need to create tons of functions

I have index, video feed, commands, auto mode, photo distance five kinds
index is a template I tested on desktop, in the end not used
commands include start, restart, take a photo, and quit, falls in the decision loop
start will initialize, check if gps and camera is off, and then turn it on, or else just refresh the view
restart will end the camera thread and open a new one by setting mp.Value from 1 to 99 (0 is rest, 1 for running, 99 for quit)
photo will first send True to the checking loop, check location is not duplicated and then send 1 to camera, after camera taken a picture will send 2 back, with log file written back to 0
auto mode is for the switch on the app, will send true/ false to the flask
drop list will send 15/25/30( still can be determined) to the camera setting script
———————————————————————————————————————–
So some things I learned in the project:
Flask:
before I did django, and while googling I found easy example with opencv + flask to stream ip cameras
mjpeg is simply read image in bytes, while in Python every thing is an object.
Socket:
socket is tcp connection in the python library, though in the end I didn’t use it, but it was a try
when sending image will first pickle the image and the size of the image
while once can only send certain bytes, to avoid corrupted data, the struct pack is needed
the codes can be seen here
Transferring from local to flask app took me some adjustment on the controls of commands between processes, and monitoring cpu usage for the little cpu of pi. It still lags a bit but its the best I can get from pi
———————————————————————————————————————–
temporary result