Python programmer to analyze raw pressure sensor data and save to Database
- or -
Post a project like this$$
- Posted:
- Proposals: 5
- Remote
- #1715618
- Expired
Description
Experience Level: Intermediate
General information for the business: OpenCV, Python
Kind of development: New program from scratch
Description of requirements/functionality: To develop a memory efficient and performant python program that employs machine learning concepts to analyze pressure sensor mat data to determine the shape of the contact surface of objects placed on the sensor mat. The program shall identify, label and track identified objects and individually save the shapes, dimensions, area, force along with a timestamp. The preferred approach is a computer vision approach that converts the pressure distribution data into a gray-scale representation that can then be analyzed using efficient computer vision packages.
-Identify the shape and dimensions of the contact surface of Items placed on the mat.
-Label and track each item detected on the mat.
-Items must be identifiable regardless of position on the mat during subsequent scans.
-Extract sum of pressure readings for each item and determine total force exerted by each item.
-Record the total force exerted by each item on each mat in the database along with a timestamp.
-The program shall be flexible enough to support various sensor densities.
-The program shall take as input the pressure distribution data.
-The program shall save data from steps 1-4 in the database and also write out was saved to the database to standard out.
Method and Approach
The analysis shall employ modern machine learning methods/algorithms as appropriate to accomplish the above tasks. The initial phase of the task shall seek to determine the most suitable algorithm for incorporation into the developed python program. Once the most efficient algorithm is determined, it shall be used in the developed program.
The file to be analyzed will be in json format:
{
"state": {
"reported":
{
"SenselID": 0,
"X": 0,
"Y": 0,
"Z": 0
},
{
"SenselID": 1,
"X": 1,
"Y": 0,
"Z": 0
},
{
"SenselID": 2,
"X": 2,
"Y": 0,
"Z": 0
}. . . . . . . . .
Based on my initial research, it appears that there are numerous packages in python that can simplify the tasks. I have posted on a number questions on ML forums about the best approach for identifying the shapes in the pressure data. There is almost unanimous consensus on an approach that one should first convert the pressure distribution data to a greyscale image representation and then using a computer vision approach to identify the shapes in the data, please see this very relevant reference: http://www.pyimagesearch.com/2016/02/08/opencv-shape-detection/
Using the approach in the reference article above, it should be trivial to identify the shapes in the pressure data file once it is converted to a greyscale image representation.
The example in this stackflow question is along the lines of what we probably need to do to get the data as a grayscale image that we can then use computer vision packages to extract shapes in the data.
https://stackoverflow.com/questions/2111150/create-a-grayscale-image
OS requirements: Linux
Extra notes: To develop a memory efficient and performant python program that employs machine learning concepts to analyze pressure sensor mat data to determine the shape of the contact surface of objects placed on the sensor mat. The program shall identify, label and track identified objects and individually save the shapes, dimensions, area, force along with a timestamp. The preferred approach is a computer vision approach that converts the pressure distribution data into a gray-scale representation that can then be analyzed using efficient computer vision packages.
Kind of development: New program from scratch
Description of requirements/functionality: To develop a memory efficient and performant python program that employs machine learning concepts to analyze pressure sensor mat data to determine the shape of the contact surface of objects placed on the sensor mat. The program shall identify, label and track identified objects and individually save the shapes, dimensions, area, force along with a timestamp. The preferred approach is a computer vision approach that converts the pressure distribution data into a gray-scale representation that can then be analyzed using efficient computer vision packages.
-Identify the shape and dimensions of the contact surface of Items placed on the mat.
-Label and track each item detected on the mat.
-Items must be identifiable regardless of position on the mat during subsequent scans.
-Extract sum of pressure readings for each item and determine total force exerted by each item.
-Record the total force exerted by each item on each mat in the database along with a timestamp.
-The program shall be flexible enough to support various sensor densities.
-The program shall take as input the pressure distribution data.
-The program shall save data from steps 1-4 in the database and also write out was saved to the database to standard out.
Method and Approach
The analysis shall employ modern machine learning methods/algorithms as appropriate to accomplish the above tasks. The initial phase of the task shall seek to determine the most suitable algorithm for incorporation into the developed python program. Once the most efficient algorithm is determined, it shall be used in the developed program.
The file to be analyzed will be in json format:
{
"state": {
"reported":
{
"SenselID": 0,
"X": 0,
"Y": 0,
"Z": 0
},
{
"SenselID": 1,
"X": 1,
"Y": 0,
"Z": 0
},
{
"SenselID": 2,
"X": 2,
"Y": 0,
"Z": 0
}. . . . . . . . .
Based on my initial research, it appears that there are numerous packages in python that can simplify the tasks. I have posted on a number questions on ML forums about the best approach for identifying the shapes in the pressure data. There is almost unanimous consensus on an approach that one should first convert the pressure distribution data to a greyscale image representation and then using a computer vision approach to identify the shapes in the data, please see this very relevant reference: http://www.pyimagesearch.com/2016/02/08/opencv-shape-detection/
Using the approach in the reference article above, it should be trivial to identify the shapes in the pressure data file once it is converted to a greyscale image representation.
The example in this stackflow question is along the lines of what we probably need to do to get the data as a grayscale image that we can then use computer vision packages to extract shapes in the data.
https://stackoverflow.com/questions/2111150/create-a-grayscale-image
OS requirements: Linux
Extra notes: To develop a memory efficient and performant python program that employs machine learning concepts to analyze pressure sensor mat data to determine the shape of the contact surface of objects placed on the sensor mat. The program shall identify, label and track identified objects and individually save the shapes, dimensions, area, force along with a timestamp. The preferred approach is a computer vision approach that converts the pressure distribution data into a gray-scale representation that can then be analyzed using efficient computer vision packages.
Ayokunle G.
0% (0)Projects Completed
-
Freelancers worked with
-
Projects awarded
0%
Last project
5 May 2024
United States
New Proposal
Login to your account and send a proposal now to get this project.
Log inClarification Board Ask a Question
-
There are no clarification messages.
We collect cookies to enable the proper functioning and security of our website, and to enhance your experience. By clicking on 'Accept All Cookies', you consent to the use of these cookies. You can change your 'Cookies Settings' at any time. For more information, please read ourCookie Policy
Cookie Settings
Accept All Cookies