| (i) | Initially load the video converted to frames in the system memory |
| (ii) | Locate the folder path P provided for the process. |
| | For every image file F belonging to the folder path P |
| | Construct file list L s.t F ∈ folder at path P |
| | Set the csv report file to empty R = []. |
| (iii) | For i ← 1 to length (L) do |
| (1) | Retrieve the image filename F(i) from L |
| (2) | Reading the image file data from F(i) into Iinp using OpenCV. |
| (3) | Annotating the images using LabelImg. |
| (4) | Creating positive data set by cropping. |
| (5) | If Iinp ∈ positive data set calculate Haar features store it in xml file. |
| (6) | Pass xml file for training the system using YOLOv4. |
| (7) | Split Iinp into S × S grid and pass it to pooling layer. |
| (8) | Outputs class probability for every bounding boxes |
| (9) | if (class prob > confidence) |
| | search bounding boxes with class label |
| | Nested if (Bounding box < class label) |
| | Go to if |
| | Else if |
| | (region > bounding box) |
| | Class = 1; //presence of vehicles and license plate |
| | Class = 0; //absence of vehicles and license plate |
| | End else if |
| | End. |
| (10) | Plot the bounding box for finding the License plate. |
| (11) | Detect the character and number by segmentation using OpenCV. |
| (12) | If (confidence <40%) |
| (13) | Nested if (number of characters are not between 8 & 10) |
| (14) | Enter the number plate into alert table |
| (15) | Else enter the number plate in non-alert table |
| (16) | Convert the alert number plates from text to speech |
| (17) | Intimation to traffic in charge for enquiry. |
