Vectorize detections

This task is used to vectorize objects detected in images.

In the input parameters it is expected:

  1. Input directory, where the images to be detected are located.

  2. Output directory, where the result of the detection will be saved in shp format.

  3. EPSG Code: The EPSG code of the input images, which will be assigned to the output SHP file.

  4. Path of the dmod file, which will be used as a model to perform the detections.

  5. Confidence threshold: Accuracy of the detections performed by the AI model.

    Only the detections with a confidence higher than the one established in this parameter will be used for the rest of the processes.

    Accuracy 0, implies that all detections will be evaluated, this may introduce erroneous detections in the process.

    Accuracy very close to 1, implies that only the clearest detections for the model will be evaluated, this may bias the detections of valid objects.

    By default, the value of 0.2 dramatically decreases anomalous detections and allows enough flexibility to maximize the number of detections.

In the advanced parameters we can configure:

  1. Number of threads: The number of parallel threads of execution.

  2. Packet processing size: The number of images that are processed in each parallel thread.

  3. Merge type: type of merging of overlapping objects of the same class, more details in the merge section.

  4. Merge intersection: Percentage of intersection by which the detections will be merged, more details in the merge section.

  5. Process by parts: This option allows the processing of the image by parts, thought for very large images.

  6. Window Width: Measurement in pixels of the width of the frame, when processing by parts.

  7. Window Height: Measure in pixels of the height of the frame, when processing by parts.

  8. Overlap Windows: Measurement in pixels, when processed in parts.

We use cookies on our website to give you the most relevant experience by remembering your preferences and repeat visits.
By clicking “Accept”, you consent to our cookies policy.