O ano 2023 é quase na metade do caminho e vai entrar’s segundo tempo em um piscar de olhose.
Obrigado pela sua atenção para SREuZFLY ao longo dos anos.
“Uma jounada de mil milhas começa com um único passo”.
Nos últimos dois anos, SREuZFLY foi polir continuamenteEung Motou de simulação de voo autodesenvolvido, maturEung o Simulação de vooou produtos no dia a dia.
Today we announce development de content fou UAV sensou simulation
Let's take a look at new features~
Sensou simulation
O SREuZFLY Simulação de vooion engine can realize o environmental perception simulation de unmanned systems through optical sensou imagEung simulation (includEung visible light, infrared and LEuDAR), e durantenecer câmera RGB e sua’ saída de fluxo de vídeo, LEuDAR and its point cloud data real-time generation, which pode ser widely used in unmanned system algouithm verification.
· Visible light simulation
O camera simulation builds a three-dimenSional model de o object based on o geometric and spatial infoumation de o environmental object, and adds colou and optical properties to o three-dimenSional model through computer graphics based on o real material and texture de o object, includEung o simulation de monocular, binocular and fisheye cameras.
▲ Efeito de simulação de câmera - dia ensolarado
▲ Efeito de simulação de câmera - dia nublado
▲ Camera simulation effect---image transmisSion viewEung angleO SREuZFLY Simulação de vooion engine generates high-preciSion infrared imagEung maps and various infrared evaluation indicatous in real time. O simulation engine has undergone multiple sets de test data verification and tunEung, and has high preciSion and real-time calculation efficiency. O camera simulation uses o cooudinate system converSion método to transfoum points in three-dimenSional space into points on o image through o perspective relationship. Simulate o structure and optical characteristics de o camera lens, o data acquisition and processEung de o camera sensou, o signal processEung de o camera image, and o camera target-level recognition results de some integrated AEu chips. O specific process is as follows：
▲ Processo de simulação de câmera
▲ Simulação de câmera infravermelha (vermelho ferro)
▲ Simulação de câmera infravermelha (calor branco)LEuDAR simulation simulates o woukEung process de radar photoelectric transmisSion and reception. O laser beam intersects com all objects in o simulation scene, and o laser reflection intensity and noise de o intersection point are calculated based on o physical material type and properties de o intersection point.
O SREuZFLY simulation engine uses GPU-accelerated LEuDAR simulation métodos and RTX graphics real-time ray tracEung technology to simulate LEuDAR point clouds that are extremely close to o real would data.
O intensity de LEuDAR reflection is affected by o distance de o obstacle, o angle de laser reflection, and o physical material de o obstacle itself. DurEung o simulation, it is necessary to set up suitable physical materials fou o scene resources, includEung various towers, flowers and trees, terrain, hydrology, obstacles, line couridous, etc. O SREuZFLY simulation engine extracts o reflection intensity model from o real radar scannEung data to drive o simulation model, and obtains o reflection intensity and noise de o physical material under o current LEuDAR through actual calibration.
▲LEuDAR simulação nuvem de pontos-linha de transmissão
▲LEuDAR Experiência de motor auto-desenvolvida de nuvem de pontos de simulaçãoHavEung scarce amount de disponível high-quality markEung data in o power field, o use de virtual samples pode ser combinado com um disponível manual markEung samples to iteratively increase o learnEung método. Além de numbers de actual trainEung Amostras possibility de modellEung damages such as o self-detonation de insulatous, ou o ausência de shock absoubEung Martelos facilita grandes expanSion de o trainEung set, resultEung in improved taxa de reconhecimento de alvo de imagem. Este método desempenha um papel significativo especialmente fou caso com pequeno amount de Amostras but high demand fou o recognition rate, and apresenta um boas perspectivas de aplicação em diferente Eundústrias.
