results_complex.qmd

---
title: "Complex tasks"
author:
  - name: Piotr Tompalski
    email: piotr.tompalski@nrcan-rncan.gc.ca
format: html
engine: knitr
lightbox: true
date: 08/15/2024
date-modified: today
code-block-bg: true
code-block-border-left: "#31BAE9"
---

The complex tasks described below consist of multiple individual steps executed within a single processing workflow, closely simulating real-world scenarios. When processing point cloud data, routines often involve a series of tasks performed sequentially, where data is read, processed, and saved at each stage. However, `lasR` allows users to build processing pipelines where the input data is read only once. This approach significantly reduces the overall processing time.

## Complex task 1

This processing routine included three tasks: generate DEM, normalize the point cloud, and generate CHM. The following `lasR` pipeline was developed:

```{r, echo=F, message=F, warning=F}
source("R/0_setup.R")
x <- benchmark_tasks %>% filter(task_id == "complex_task1") %>% pull(task_call_lasR)
x <- str_remove_all(string = x, pattern = "    ")
cat(x)

```
The total processing time for data processed sequentially was calculated by summing the processing times of the three basic tasks included in the routine.

### Results

![](graphics/result_tasks_complex1.png)


![](graphics/result_tasks_complex1_delta.png)