{"id":3900,"date":"2021-12-21T14:00:38","date_gmt":"2021-12-21T22:00:38","guid":{"rendered":"https:\/\/www.cloudacm.com\/?p=3900"},"modified":"2021-12-13T07:12:51","modified_gmt":"2021-12-13T15:12:51","slug":"python-data-visualization","status":"publish","type":"post","link":"https:\/\/www.cloudacm.com\/?p=3900","title":{"rendered":"Python Data Visualization"},"content":{"rendered":"

In an earlier post, python was used to generate an image from datasets. That method lacked the full features needed to create a final image. This post will introduce Plotly as a preferred method. It offers more options and is targeted for use as a tool for dataset visualization.<\/p>\n

First we’ll need a dataset that is tangible. In this post the dataset will be generated from a field scan to create an image array. The device used to scan the field will be a photoresistor that will be pivoted by a pair of servos, each one on a x and y axis. This will be purely a theoretical scenario, so no actual code or hardware was developed for this concept.<\/p>\n

The photoresistor used will be a standard cadmium sulfide LDR. The sensor was fitted with a focal guide to allow only a narrow angular field of view. This was done by fitting heat shrink tubing around the sensor and allowing a long narrow opening for light to enter. The datasheet can be found online here, https:\/\/www.tme.eu\/Document\/0b7aec6d26675b47f9e54d893cd4521b\/PGM5506.pdf<\/a><\/p>\n

As mentioned, the photoresistor sensor will take stepped measurements of a field horizontally and continue that pattern for each step vertically . The step of each motion has been set to 1 degree and is centered on the photoresistor’s sensor surface. Likewise, the pivot on the y axis is also centered on the photoresistor’s sensor surface with each motion step set to 1 degree. The range of motion will be limited to 60 degrees on the x axis and 35 degrees on the y. This should result in an array that has a 60 x 35 pixel resolution.<\/p>\n

The speed that the servos can step will be limited by the response time of the photoresistor as defined in its datasheet. The larger value from either the rise or decay time for the sensor will determine dwell time for each measurement step or sample. Based on the datasheet, a dwell time of 50ms should be enough for the sensor to acquire its reading. Based on that limit, the entire time to scan the field will be x * y * 50 = time in ms, which is 1 minute and 45 seconds. So the object and photoresistor must be stationary for a reliable image to be created, which is why a still life was selected.<\/p>\n

Any microcontroller that can be programmed to step the pair of servos, quantify a value measured by the sensor, format the readings with delimitation and carriage returns, and store the reading to memory is suitable for use. The final results should look similar to this.<\/p>\n

135,135,135,....159,159,159\r\n136,136,136,....159,159,159\r\n...\r\n138,137,137,....160,160,161<\/pre>\n
The dataset source file is here, SCN_IMG_GRAY_Array1<\/a><\/p>\n
Once the scan is completed, the dataset can be processed by Plotly to create the image. Here is the python code used with this command “python3 SCN_IMG_GRAY_Array1_Plotly.py”<\/p>\n
import plotly.express as px\r\nfrom numpy import genfromtxt\r\n\r\ndata = genfromtxt('SCN_IMG_GRAY_Array1_Plotly.csv', delimiter=',')\r\n\r\nfig = px.imshow(data, color_continuous_scale=px.colors.sequential.gray)\r\n                 \r\nfig.update_layout(width=60, height=35, margin=dict(l=0, r=0, b=0, t=0), coloraxis_showscale=False)\r\nfig.update_xaxes(showticklabels=False).update_yaxes(showticklabels=False)\r\n# fig.show()\r\n\r\nfig.write_image(\"SCN_IMG_GRAY_Array1_Plotly.png\") \r\n\r\n# Command - python3 SCN_IMG_GRAY_Array1_Plotly.py\r\n# Source - https:\/\/plotly.com\/python\/heatmaps\/\r\n# Color Scales - https:\/\/plotly.com\/python\/builtin-colorscales\/\r\n\r\n# https:\/\/plotly.com\/python\/imshow\/\r\n# https:\/\/pypi.org\/project\/plotly-express\/\r\n# https:\/\/www.cloudacm.com\/?p=2674\r\n# https:\/\/plotly.com\/python\/plotly-express\/<\/pre>\n
 <\/p>\n
Here is the resulting image file generated, which has been zoomed 10 times in order to allow for easier viewing.<\/p>\n
\"\"<\/a><\/p>\n
It is remarkable that a simple photoresistor can create a grayscale image, but with the use of color filters, datasets can be merged to create a full color image. This has been demontrated here with some background on the process.<\/p>\n