Trackers are now generic over the Blob[Self] interface, enabling custom blob implementations.
Before:
tracker := mot.NewNewSimpleTracker(15.0, 5)After:
tracker := mot.NewNewSimpleTracker[*mot.SimpleBlob](15.0, 5)This change enables future implementations like BlobBBox (Kalman filter tracking with full bounding box dynamics) without duplicating tracker code.
This is my second attempt to create some pretty naive implementation of multi-object tracker. The first one was gocv-blob.
This one uses similar approach to this implementation in Rust programming language.
What is this good for?
You can use this library to track vehicles / peoples and etc. when you don't need that much accuracy or ReID.
What Multi-Object tracking algorithms are implemented?
- Centroids distance + diagonal - mot/simple_tracker#38
- ByteTrack using greedy matching algorithm - mot/bytetrack.go#266
- ByteTrack using Hungarian algorithm via go-hungarian package - mot/bytetrack.go#231
Are more advanced algorithms considered to be implemented in futher?
Yes, I do think so. I guess that SORT or naive IoU tracker will be the next one.
If you want to you can contribute via opening Pull Request
Some examples
| Simple centroid IoU tracker for three simple tracks | ByteTrack + Hungarian algorithm for three simple tracks |
|---|---|
 |
 |
| Simple centroid IoU tracker for spread tracks | ByteTrack + Hungarian algorithm for spread tracks |
|---|---|
 |
 |
Add dependency into your project:
go get github.com/LdDl/mot-goLet's create really synthetic example and define similar trajectories for three objects. We're using pretty simple MOT algorithm, so no hard tasks for now.
package main
import (
"encoding/csv"
"fmt"
"os"
"strings"
"github.com/LdDl/mot-go/mot"
)
func main() {
bboxesOne := [][]float64{[]float64{236, -25, 386, 35}, []float64{237, -24, 387, 36}, []float64{238, -22, 388, 38}, []float64{236, -20, 386, 40}, []float64{236, -19, 386, 41}, []float64{237, -18, 387, 42}, []float64{237, -18, 387, 42}, []float64{238, -17, 388, 43}, []float64{237, -14, 387, 46}, []float64{237, -14, 387, 46}, []float64{237, -12, 387, 48}, []float64{237, -12, 387, 48}, []float64{237, -11, 387, 49}, []float64{237, -11, 387, 49}, []float64{237, -10, 387, 50}, []float64{237, -10, 387, 50}, []float64{237, -8, 387, 52}, []float64{237, -8, 387, 52}, []float64{236, -7, 386, 53}, []float64{236, -7, 386, 53}, []float64{236, -6, 386, 54}, []float64{236, -6, 386, 54}, []float64{236, -2, 386, 58}, []float64{235, 0, 385, 60}, []float64{236, 2, 386, 62}, []float64{236, 5, 386, 65}, []float64{236, 9, 386, 69}, []float64{235, 12, 385, 72}, []float64{235, 14, 385, 74}, []float64{233, 16, 383, 76}, []float64{232, 26, 382, 86}, []float64{233, 28, 383, 88}, []float64{233, 40, 383, 100}, []float64{233, 30, 383, 90}, []float64{232, 22, 382, 82}, []float64{232, 34, 382, 94}, []float64{232, 21, 382, 81}, []float64{233, 40, 383, 100}, []float64{232, 40, 382, 100}, []float64{232, 40, 382, 100}, []float64{232, 36, 382, 96}, []float64{232, 53, 382, 113}, []float64{232, 50, 382, 110}, []float64{233, 55, 383, 115}, []float64{232, 50, 382, 110}, []float64{234, 68, 384, 128}, []float64{231, 49, 381, 109}, []float64{232, 68, 382, 128}, []float64{231, 31, 381, 91}, []float64{232, 64, 382, 124}, []float64{233, 71, 383, 131}, []float64{231, 64, 381, 124}, []float64{231, 74, 381, 134}, []float64{231, 64, 381, 124}, []float64{230, 77, 380, 137}, []float64{232, 82, 382, 142}, []float64{232, 78, 382, 138}, []float64{232, 78, 382, 138}, []float64{231, 79, 381, 139}, []float64{231, 79, 381, 139}, []float64{231, 91, 381, 151}, []float64{232, 78, 382, 138}, []float64{232, 78, 382, 138}, []float64{233, 90, 383, 150}, []float64{232, 92, 382, 152}, []float64{232, 92, 382, 152}, []float64{233, 98, 383, 158}, []float64{232, 100, 382, 160}, []float64{231, 92, 381, 152}, []float64{233, 110, 383, 170}, []float64{234, 92, 384, 152}, []float64{234, 92, 384, 152}, []float64{234, 110, 384, 170}, []float64{234, 92, 384, 152}, []float64{233, 104, 383, 164}, []float64{234, 111, 384, 171}, []float64{234, 106, 384, 166}, []float64{234, 106, 384, 166}, []float64{233, 124, 383, 184}, []float64{236, 125, 386, 185}, []float64{236, 125, 386, 185}, []float64{232, 120, 382, 180}, []float64{236, 131, 386, 191}, []float64{232, 132, 382, 192}, []float64{238, 139, 388, 199}, []float64{236, 141, 386, 201}, []float64{232, 151, 382, 211}, []float64{236, 145, 386, 205}, []float64{236, 145, 386, 205}, []float64{231, 133, 381, 193}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}, []float64{237, 148, 387, 208}}
bboxesTwo := [][]float64{[]float64{321, -25, 471, 35}, []float64{322, -24, 472, 36}, []float64{323, -22, 473, 38}, []float64{321, -20, 471, 40}, []float64{321, -19, 471, 41}, []float64{322, -18, 472, 42}, []float64{322, -18, 472, 42}, []float64{323, -17, 473, 43}, []float64{322, -14, 472, 46}, []float64{322, -14, 472, 46}, []float64{322, -12, 472, 48}, []float64{322, -12, 472, 48}, []float64{322, -11, 472, 49}, []float64{322, -11, 472, 49}, []float64{322, -10, 472, 50}, []float64{322, -10, 472, 50}, []float64{322, -8, 472, 52}, []float64{322, -8, 472, 52}, []float64{321, -7, 471, 53}, []float64{321, -7, 471, 53}, []float64{321, -6, 471, 54}, []float64{321, -6, 471, 54}, []float64{321, -2, 471, 58}, []float64{320, 0, 470, 60}, []float64{321, 2, 471, 62}, []float64{321, 5, 471, 65}, []float64{321, 9, 471, 69}, []float64{320, 12, 470, 72}, []float64{320, 14, 470, 74}, []float64{318, 16, 468, 76}, []float64{317, 26, 467, 86}, []float64{318, 28, 468, 88}, []float64{318, 40, 468, 100}, []float64{318, 30, 468, 90}, []float64{317, 22, 467, 82}, []float64{317, 34, 467, 94}, []float64{317, 21, 467, 81}, []float64{318, 40, 468, 100}, []float64{317, 40, 467, 100}, []float64{317, 40, 467, 100}, []float64{317, 36, 467, 96}, []float64{317, 53, 467, 113}, []float64{317, 50, 467, 110}, []float64{318, 55, 468, 115}, []float64{317, 50, 467, 110}, []float64{319, 68, 469, 128}, []float64{316, 49, 466, 109}, []float64{317, 68, 467, 128}, []float64{316, 31, 466, 91}, []float64{317, 64, 467, 124}, []float64{318, 71, 468, 131}, []float64{316, 64, 466, 124}, []float64{316, 74, 466, 134}, []float64{316, 64, 466, 124}, []float64{315, 77, 465, 137}, []float64{317, 82, 467, 142}, []float64{317, 78, 467, 138}, []float64{317, 78, 467, 138}, []float64{316, 79, 466, 139}, []float64{316, 79, 466, 139}, []float64{316, 91, 466, 151}, []float64{317, 78, 467, 138}, []float64{317, 78, 467, 138}, []float64{318, 90, 468, 150}, []float64{317, 92, 467, 152}, []float64{317, 92, 467, 152}, []float64{318, 98, 468, 158}, []float64{317, 100, 467, 160}, []float64{316, 92, 466, 152}, []float64{318, 110, 468, 170}, []float64{319, 92, 469, 152}, []float64{319, 92, 469, 152}, []float64{319, 110, 469, 170}, []float64{319, 92, 469, 152}, []float64{318, 104, 468, 164}, []float64{319, 111, 469, 171}, []float64{319, 106, 469, 166}, []float64{319, 106, 469, 166}, []float64{318, 124, 468, 184}, []float64{321, 125, 471, 185}, []float64{321, 125, 471, 185}, []float64{317, 120, 467, 180}, []float64{321, 131, 471, 191}, []float64{317, 132, 467, 192}, []float64{323, 139, 473, 199}, []float64{321, 141, 471, 201}, []float64{317, 151, 467, 211}, []float64{321, 145, 471, 205}, []float64{321, 145, 471, 205}, []float64{316, 133, 466, 193}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}, []float64{322, 148, 472, 208}}
bboxesThree := [][]float64{[]float64{151, -25, 301, 35}, []float64{152, -24, 302, 36}, []float64{153, -22, 303, 38}, []float64{151, -20, 301, 40}, []float64{151, -19, 301, 41}, []float64{152, -18, 302, 42}, []float64{152, -18, 302, 42}, []float64{153, -17, 303, 43}, []float64{152, -14, 302, 46}, []float64{152, -14, 302, 46}, []float64{152, -12, 302, 48}, []float64{152, -12, 302, 48}, []float64{152, -11, 302, 49}, []float64{152, -11, 302, 49}, []float64{152, -10, 302, 50}, []float64{152, -10, 302, 50}, []float64{152, -8, 302, 52}, []float64{152, -8, 302, 52}, []float64{151, -7, 301, 53}, []float64{151, -7, 301, 53}, []float64{151, -6, 301, 54}, []float64{151, -6, 301, 54}, []float64{151, -2, 301, 58}, []float64{150, 0, 300, 60}, []float64{151, 2, 301, 62}, []float64{151, 5, 301, 65}, []float64{151, 9, 301, 69}, []float64{150, 12, 300, 72}, []float64{150, 14, 300, 74}, []float64{148, 16, 298, 76}, []float64{147, 26, 297, 86}, []float64{148, 28, 298, 88}, []float64{148, 40, 298, 100}, []float64{148, 30, 298, 90}, []float64{147, 22, 297, 82}, []float64{147, 34, 297, 94}, []float64{147, 21, 297, 81}, []float64{148, 40, 298, 100}, []float64{147, 40, 297, 100}, []float64{147, 40, 297, 100}, []float64{147, 36, 297, 96}, []float64{147, 53, 297, 113}, []float64{147, 50, 297, 110}, []float64{148, 55, 298, 115}, []float64{147, 50, 297, 110}, []float64{149, 68, 299, 128}, []float64{146, 49, 296, 109}, []float64{147, 68, 297, 128}, []float64{146, 31, 296, 91}, []float64{147, 64, 297, 124}, []float64{148, 71, 298, 131}, []float64{146, 64, 296, 124}, []float64{146, 74, 296, 134}, []float64{146, 64, 296, 124}, []float64{145, 77, 295, 137}, []float64{147, 82, 297, 142}, []float64{147, 78, 297, 138}, []float64{147, 78, 297, 138}, []float64{146, 79, 296, 139}, []float64{146, 79, 296, 139}, []float64{146, 91, 296, 151}, []float64{147, 78, 297, 138}, []float64{147, 78, 297, 138}, []float64{148, 90, 298, 150}, []float64{147, 92, 297, 152}, []float64{147, 92, 297, 152}, []float64{148, 98, 298, 158}, []float64{147, 100, 297, 160}, []float64{146, 92, 296, 152}, []float64{148, 110, 298, 170}, []float64{149, 92, 299, 152}, []float64{149, 92, 299, 152}, []float64{149, 110, 299, 170}, []float64{149, 92, 299, 152}, []float64{148, 104, 298, 164}, []float64{149, 111, 299, 171}, []float64{149, 106, 299, 166}, []float64{149, 106, 299, 166}, []float64{148, 124, 298, 184}, []float64{151, 125, 301, 185}, []float64{151, 125, 301, 185}, []float64{147, 120, 297, 180}, []float64{151, 131, 301, 191}, []float64{147, 132, 297, 192}, []float64{153, 139, 303, 199}, []float64{151, 141, 301, 201}, []float64{147, 151, 297, 211}, []float64{151, 145, 301, 205}, []float64{151, 145, 301, 205}, []float64{146, 133, 296, 193}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}, []float64{152, 148, 302, 208}}
tracker := mot.NewNewSimpleTracker[*mot.SimpleBlob](15.0, 5)
dt := 1.0 / 25.0 // emulate 25 fps
for idx := range bboxesOne {
rectOne := mot.NewRect(bboxesOne[idx][0], bboxesOne[idx][1], bboxesOne[idx][2]-bboxesOne[idx][0], bboxesOne[idx][3]-bboxesOne[idx][1])
rectTwo := mot.NewRect(bboxesTwo[idx][0], bboxesTwo[idx][1], bboxesTwo[idx][2]-bboxesTwo[idx][0], bboxesTwo[idx][3]-bboxesTwo[idx][1])
rectThree := mot.NewRect(bboxesThree[idx][0], bboxesThree[idx][1], bboxesThree[idx][2]-bboxesThree[idx][0], bboxesThree[idx][3]-bboxesThree[idx][1])
blobOne := mot.NewSimpleBlobWithTime(rectOne, dt)
blobTwo := mot.NewSimpleBlobWithTime(rectTwo, dt)
blobThree := mot.NewSimpleBlobWithTime(rectThree, dt)
blobs := []*mot.SimpleBlob{blobOne, blobTwo, blobThree}
err := tracker.MatchObjects(blobs)
if err != nil {
fmt.Println(err)
return
}
}
file, err := os.Create("blobs_naive.csv")
if err != nil {
fmt.Println(err)
return
}
defer file.Close()
writer := csv.NewWriter(file)
defer writer.Flush()
writer.Comma = ';'
err = writer.Write([]string{"id", "track"})
if err != nil {
fmt.Println(err)
return
}
for objectID, object := range tracker.Objects {
track := object.GetTrack()
data := make([]string, len(track))
for idx, pt := range track {
data[idx] = fmt.Sprintf("%f,%f", pt.X, pt.Y)
}
dataStr := strings.Join(data, "|")
err = writer.Write([]string{objectID.String(), dataStr})
if err != nil {
fmt.Println(err)
return
}
}
}If we plot results of filtered tracks we should get something like:
| Similar (for code example above) | Spread |
|---|---|
|
|
Trajectories